Topics for master theses

Topics for master theses

The Department of Business and Management Science can offer the following Master Theses topics:

Investments, Insurance and Household Finance

  • Performance evaluation for mutual funds

    Performance evaluation for mutual funds

    Supervisor: Trond M. Døskeland.

  • Identifying financial behavioral biases for households

    Identifying financial behavioral biases for households

    Supervisor: Trond M. Døskeland.

  • Risk Management

    Risk Management

    Analyze a given company or a given line of business exposed to price risk from commodities/services sold or bought. May also be combined with currency risk. What is the level of risk, how do we measure it, which instruments are available to deal with this risk and how successful is the risk-reducing strategy, using these instruments. Hedging strategies are based on a combination of risk-reduction and expected values, so both components may be included in the analyses.

    Supervisor: Øystein Gjerde.

  • How did the US Tax Bill affect M&As?

    How did the US Tax Bill affect M&As?

    The 2017 TCJA may significantly change the pattern of corporate M&As. For instance, it has lifted the original use requirement for immediate depreciation of acquired assets. That makes it much more attractive to structure deals as asset purchases rather than stock acquisitions.

    Furthermore, net operating losses before and after January 1, 2018 are treated differently. An empirical analysis could use either Zephyr or SDC Platinum to explore early signs of these changes. 

    Supervisor: Maximilian Todtenhaupt.

  • Common ownership and tax avoidance

    Common ownership and tax avoidance

    Corporate tax planning is high on the agenda of both tax practitioners and policy makers. While some firms certainly try to avoid paying taxes, we know very little about how firms learn about different tax avoidance mechanisms. An important channel may be via common owners. Do firms engage in similar levels of tax avoidance if they are owned by the same investors? This project will investigate this question using balance sheet information and ownership details of large corporations.

    Supervisor: Maximilian Todtenhaupt.

  • Shipping & Finance: How do risk capital and risk limits affect the chartering policy of a ship operator?

    Shipping & Finance: How do risk capital and risk limits affect the chartering policy of a ship operator?

    Using: freight rates timeseries, optimal portfolio theory.

    Company: Western Bulk.

    Supervisor: Professor Roar Ådland

  • Shipping & Finance: Machine learning models for FFA trading

    Shipping & Finance: Machine learning models for FFA trading

    Using spatial AIS data for ship positions, open-source weather and macro data – can you develop a machine learning model to generate profitable trading signals? Requires knowledge of Python, implementation of machine learning models.

    Supervisor: Roar Ådland and Haiying Jia

  • Shipping & Finance: Inferring short-term market direction from intraday FFA Data

    Shipping & Finance: Inferring short-term market direction from intraday FFA Data

    Using intraday bid/offers spreads – is it possible to use pattern recognition or technical analysis to daytrade forward freight agreements? Knowledge of Python and machine learning is useful. Can possibly be done in cooperation with Clarksons.

    Supervisor: Professor Roar Ådland

  • Risk modelling in ship finance and marine insurance

    Risk modelling in ship finance and marine insurance

    How to aggregate risk exposure in the content of credit risk, geographical exposure, and event risks etc.? How to optimally transfer risks in reinsurance? How to evaluate financial returns with regards to low probability - high severity risks?

    Supervisor: Haiying Jia

  • Sustainable ship finance

    Sustainable ship finance

    How ship finance helps to achieve green shipping, using Poseidon Principles as a case study.

    Supervisor: Haiying Jia

  • Alternative shipping capital sources

    Alternative shipping capital sources

    The development of Chinese shipping trusts as alternative shipping finance sources.

    Supervisor: Haiying Jia

  • Blockchain and smart contracts in shipping and supply chain management

    Blockchain and smart contracts in shipping and supply chain management

    How will blockchain change the relationships among shipping players? Which areas do blockchain provide the most value for the shipping ecosystem? May be survey based. Incentive mechanism in a blockchain-based smart contract in shipping - how to incentivize players to participate?

    Supervisor: Haiying Jia

  • Finansiell økonomi (finansmarkeder/opsjoner/skatt): Er leterefusjonsordningen subsidiering?

    Finansiell økonomi (finansmarkeder/opsjoner/skatt): Er leterefusjonsordningen subsidiering?

    Den norske leterefusjonsordningen har i det siste vært under offentlig debatt. Ordningen går ut på at oljeselskaper som er utenfor skatteposisjon får skatterefusjon for leteutgifter. Ifølge myndighetene innebærer ordningen ikke subsidiering fordi den gir likebehandling mellom selskaper som er utenfor skatteposisjon og selskaper som er i skatteposisjon.

    Utfordring: Belyse denne problemstillingen ut fra økonomisk teori.

    Se: Kjell-Børge Freiberg og Siv Jensen: «Oljeleting gir millarder til statskassen». Innlegg i Dagens Næringsliv, 8. februar 2019.

    https://www.dn.no/innlegg/okonomi/energi/leterefusjonsordningen/oljeleting-gir-milliarder-til-statskassen/2-1-537307

    Supervisor: Petter Bjerksund.

  • Finansiell økonomi (kapitalforvaltning/personlig økonomi): Seniorlån – en melkeku for bankene på arvingenes bekostning?

    Finansiell økonomi (kapitalforvaltning/personlig økonomi): Seniorlån – en melkeku for bankene på arvingenes bekostning?

    Aktuelle spørsmål:

    1. Hvordan er lånene priset?
    2. Hvordan er konkurransen i dette markedet?
    3. Hva er praksis internasjonalt?
    4. Er det greit å friste til forbruk som går på arven løs?

    https://www.bnbank.no/person/lan/seniorlan/

    https://www.klp.no/bank-og-lan/vare-boliglan/seniorlan

    Supervisor: Petter Bjerksund.

  • Using Textual Analysis to identify whether there is a gender gap in financial white collar crime

    Using Textual Analysis to identify whether there is a gender gap in financial white collar crime

    There is very little systematic evidence on the gender gap in crime. In particular, it is difficult to quantify it for financial white collar crime, as they do not find their way into police statistics and as there is little female presence on the top levels of companies.

    In this project, we can use textual analysis tools to collect data from the Securities and Exchange Commission (SEC) on insider trading and other frauds. We can match the name of the defendant to a gender, and quantify what is the difference between males and females.

    Then, we can correlate the gap, as well as the fraud itself, to past financial statements of involved companies in terms of gender representation in the board of the company and other indicators of company culture. Finally, we can compare the gap to other measures of female participation in the boardroom and determine whether white collar females seem more or less likely to commit crimes than white collar males. We can provide a partial answer to the question: Are companies going to become more responsible (do less criminal rule-breaking) if there are more females on the board?

    Methods: Textual analysis, web crawling, R

    Supervisor: Evelina Gavrilova-Zoutman.

  • What do companies hide? Geographic disclosures, tax haven presence and earnings statements

    What do companies hide? Geographic disclosures, tax haven presence and earnings statements

    Using textual analysis on earnings statements from the Securities and Exchange Commission EDGAR database, we can determine what do companies disclose about their operations in tax havens.

    On one hand, it is common knowledge that companies, who have affiliates in tax havens, do not actually conduct any real economic activity in these tax havens. They use these affiliates to avoid taxation through profit shifting. Through textual analysis, we can determine in what context are tax havens mentioned in financial statements and whether there is real activity associated with them. We can compare that to actual company statements and determine whether there is a difference between claims and reality.

    Do companies discriminate between tax havens and secrecy havens? Are they more likely to name specifically secrecy jurisdictions in the geographic disclosure, because they know that these countries would never reveal any information? Or, would the naming of such an affiliate be a bad signal for the presence of shady activities? Has this affiliate been named in press releases, if yes, with what sentiment?

    Methods: Textual analysis, web crawling, R

    Starting point: H Akamah, OK Hope, WB Thomas - Journal of International Business Studies, 2018

    Supervisor: Evelina Gavrilova-Zoutman.

  • Dividend taxation, abusive stock swap and loan transactions

    Dividend taxation, abusive stock swap and loan transactions

    A tax strategy among investors is to recast a dividend payment as a swap payment in order to take advantage of favoured tax treatment given to swap agreements involving non-US persons in the US. US stock dividents paid to non-US persons are subject to the dividend tax, whereas “dividend equivalents” paid to non-US persons as part of a swap agreement are not subject to any US tax.

    Since a 2009 Senate hearing identified the tax evasion nature of these transactions, there has been little research on the topic. These transactions are likely the origin from which cum fraud schemes have arisen. Since 2009 the US has had reforms in their dividend tax and in the legislation surrounding the claiming of these dividend equivalents. What is the impact of these reforms on stock lending of US stocks?

    Supervisor: Floris Zoutman and Evelina Gavrilova-Zoutman

  • Dividend taxation, Cum-Cum Trading and Ex-Dividend Pricing

    Dividend taxation, Cum-Cum Trading and Ex-Dividend Pricing

    A common tax planning strategy among investors is to sell their stocks the day before dividends are due, and buy them back on the ex-dividend day. This strategy, known as cum-cum trading, allows investors to avoid paying dividend taxes. The Norwegian tax authorities are considering to implement policies that make cum-cum trading less attractive in order to generate more dividend tax revenue.

    Your task will be to see how cum-cum trading relates to taxation and other policy variables, using international stock market data. Questions that you could answer in this topic are: Do stock market experience excess trade around the dividend day? Does excess trade relate to the dividend tax rate? Do stock prices reflect the level of the dividend tax? Are policies aimed at combatting cum-cum trading effective in other countries?

    Cum-cum trading strategies have been detected in Europe. Are they present in the Asian and South-American market? Are these trades a global problem that contributes to rising inequality?

    Supervisor: Floris Zoutman and Evelina Gavrilova-Zoutman

Business Taxation

  • Dividend taxation, Cum-Cum Trading and Ex-Dividend Pricing

    Dividend taxation, Cum-Cum Trading and Ex-Dividend Pricing

    A common tax planning strategy among investors is to sell their stocks the day before dividends are due, and buy them back on the ex-dividend day. This strategy, known as cum-cum trading, allows investors to avoid paying dividend taxes. The Norwegian tax authorities are considering to implement policies that make cum-cum trading less attractive in order to generate more dividend tax revenue.

    Your task will be to see how cum-cum trading relates to taxation and other policy variables, using international stock market data. Questions that you could answer in this topic are: Do stock market experience excess trade around the dividend day? Does excess trade relate to the dividend tax rate? Do stock prices reflect the level of the dividend tax? Are policies aimed at combatting cum-cum trading effective in other countries?

    Cum-cum trading strategies have been detected in Europe. Are they present in the Asian and South-American market? Are these trades a global problem that contributes to rising inequality?

    Supervisor: Floris Zoutman and Evelina Gavrilova-Zoutman

  • Dividend taxation, abusive stock swap and loan transactions

    Dividend taxation, abusive stock swap and loan transactions

    A tax strategy among investors is to recast a dividend payment as a swap payment in order to take advantage of favoured tax treatment given to swap agreements involving non-US persons in the US. US stock dividents paid to non-US persons are subject to the dividend tax, whereas “dividend equivalents” paid to non-US persons as part of a swap agreement are not subject to any US tax.

    Since a 2009 Senate hearing identified the tax evasion nature of these transactions, there has been little research on the topic. These transactions are likely the origin from which cum fraud schemes have arisen. Since 2009 the US has had reforms in their dividend tax and in the legislation surrounding the claiming of these dividend equivalents. What is the impact of these reforms on stock lending of US stocks?

    Supervisor: Floris Zoutman and Evelina Gavrilova-Zoutman

  • How did the US Tax Bill affect M&As?

    How did the US Tax Bill affect M&As?

    The 2017 TCJA may significantly change the pattern of corporate M&As. For instance, it has lifted the original use requirement for immediate depreciation of acquired assets. That makes it much more attractive to structure deals as asset purchases rather than stock acquisitions.

    Furthermore, net operating losses before and after January 1, 2018 are treated differently. An empirical analysis could use either Zephyr or SDC Platinum to explore early signs of these changes. 

    Supervisor: Maximilian Todtenhaupt.

  • Common ownership and tax avoidance

    Common ownership and tax avoidance

    Corporate tax planning is high on the agenda of both tax practitioners and policy makers. While some firms certainly try to avoid paying taxes, we know very little about how firms learn about different tax avoidance mechanisms. An important channel may be via common owners. Do firms engage in similar levels of tax avoidance if they are owned by the same investors? This project will investigate this question using balance sheet information and ownership details of large corporations.

    Supervisor: Maximilian Todtenhaupt.

  • Inventors and tax havens

    Inventors and tax havens

    Inventors are an important source of innovation for any country. At the same time they are highly mobile and respond to tax incentives. Furthermore, the intellectual property they create (e.g. patents) can be used shift income to tax havens. How many inventors are involved in such tax avoidance behavior? This project will assess the importance of inventors in tax havens by combining data on international inventors with the Panama papers which have recently become available.

    Supervisor: Maximilian Todtenhaupt.

  • The Tax Haven Call

    The Tax Haven Call

    In political and institutional economy we think of countries as having extractive and inclusive institutions. Institutions are loosely defined as informal norms of behaviour. In an influential (but controversial) paper by Acemoglu et al. (2001) the type of institutions are shown to impact economic development. However, it is unclear whether what aspect of these informal norms have influenced countries like the Netherlands, Ireland and Bermuda to become tax havens?

    Dharmapala and Hines (2009) have found that governance is an important factor that separates tax havens from non-tax havens. Better-governed small countries are more likely to be successful tax havens than badly governed small countries. Governance and institutions are closely related, but the link is not explored in this article.

    In addition, the list of tax havens has expanded since 2009 and now we have continuous measures of secrecy and tax haven status, which can give better identification in re-examining the question: What makes a tax haven?

    Starting point: Dharmapala, D. and Hines Jr, J.R., 2009. Which countries become tax havens?. Journal of Public Economics, 93(9-10), pp.1058-1068.

    Supervisor: Evelina Gavrilova-Zoutman.

  • Wealth Taxation in Norway

    Wealth Taxation in Norway

    The wealth tax is one of the most controversial aspects of the Norwegian tax system. Detractors believe that the wealth tax hurts economic growth by disincentivizing entrepreneurial activity and risk taking, and taking away a source of liquidity for business owners. Proponents argue that the wealth tax is a great instrument to reduce inequality.

    In this project you will use data on Norwegian tax payers to evaluate the arguments of the detractors. Specifically, the goal is to understand i.) whether the arguments are valid empirically, ii.) how large the concerns are quantitatively. Specific research questions could be i.) does the wealth tax discourage individuals/business owners from taking risk, ii.) does the wealth tax reduce liquidity for small businesses, iii.) does the wealth tax reduce innovation or iv). does the wealth tax discourage savings. The answer of each of these four questions is of great practical relevance to policy makers that have to make a trade-off between the efficiency cost and the equity gain associated with the wealth tax.

    Data: Individual tax return data

    Literature:

    • Berzins, Janis, Øyvind Bøhren and Bodan Stacesu (2019). Illiquid Shareholders and real firm effects: the personal wealth tax and financial constraints. Working Paper BI.
    • Akcigit, U., Grigsby, J., Nicholas, T., & Stantcheva, S. (2018). Taxation and Innovation in the 20th Century. NBER Working Paper.

    Supervisor: Floris Zoutman

  • Taxation and Innovation

    Taxation and Innovation

    In a recent working paper Akcigit et al (2018) use US data to show that higher income tax rates reduce innovation. The aim of this thesis is to find out if a similar relationship between innovation and taxation exists for Europe. Students will use the universe of patent applications to the European patent office to find out i.) if countries with lower tax rates file for more patents and ii.) whether changes in the tax rate cause changes in patent applications. The outcome is particularly relevant for policy makers, since innovation is an important driver of economic growth. If higher tax rates indeed reduce innovation, this provides a strong incentive for countries not to increase their tax rate, or even to reduce it.

    Data: Patent applications in Europe

    Literature: Akcigit, U., Grigsby, J., Nicholas, T., & Stantcheva, S. (2018). Taxation and Innovation in the 20th Century. NBER Working Paper.

    Supervisor: Floris Zoutman and Steffen Juranek

  • What do companies hide? Geographic disclosures, tax haven presence and earnings statements

    What do companies hide? Geographic disclosures, tax haven presence and earnings statements

    Using textual analysis on earnings statements from the Securities and Exchange Commission EDGAR database, we can determine what do companies disclose about their operations in tax havens.

    On one hand, it is common knowledge that companies, who have affiliates in tax havens, do not actually conduct any real economic activity in these tax havens. They use these affiliates to avoid taxation through profit shifting. Through textual analysis, we can determine in what context are tax havens mentioned in financial statements and whether there is real activity associated with them. We can compare that to actual company statements and determine whether there is a difference between claims and reality.

    Do companies discriminate between tax havens and secrecy havens? Are they more likely to name specifically secrecy jurisdictions in the geographic disclosure, because they know that these countries would never reveal any information? Or, would the naming of such an affiliate be a bad signal for the presence of shady activities? Has this affiliate been named in press releases, if yes, with what sentiment?

    Methods: Textual analysis, web crawling, R

    Starting point: H Akamah, OK Hope, WB Thomas - Journal of International Business Studies, 2018

    Supervisor: Evelina Gavrilova-Zoutman.

  • Using Textual Analysis to identify whether there is a gender gap in financial white collar crime

    Using Textual Analysis to identify whether there is a gender gap in financial white collar crime

    There is very little systematic evidence on the gender gap in crime. In particular, it is difficult to quantify it for financial white collar crime, as they do not find their way into police statistics and as there is little female presence on the top levels of companies.

    In this project, we can use textual analysis tools to collect data from the Securities and Exchange Commission (SEC) on insider trading and other frauds. We can match the name of the defendant to a gender, and quantify what is the difference between males and females.

    Then, we can correlate the gap, as well as the fraud itself, to past financial statements of involved companies in terms of gender representation in the board of the company and other indicators of company culture. Finally, we can compare the gap to other measures of female participation in the boardroom and determine whether white collar females seem more or less likely to commit crimes than white collar males. We can provide a partial answer to the question: Are companies going to become more responsible (do less criminal rule-breaking) if there are more females on the board?

    Methods: Textual analysis, web crawling, R

    Supervisor: Evelina Gavrilova-Zoutman.

  • Who blows the whistle on cartels?

    Who blows the whistle on cartels?

    By using webcrawling and textual analysis, you can gather data from the EU commission website on who reports the presence of market cartels. What is the reward that these whistleblowers get? Did they participate in the cartel before they reported it? Are these companies characterized by better corporate governance? Why did they blow the whistle? Could it be because of big shifts in governance structure, new boards, etc? What are sentiments on social networks on the mentions of the cartels and on the activity of the whistleblower?

    Are cartels more likely to be reported in certain industries? By using the revealed data on these cartels, how can we tailor a detection strategy that would detect similar behaviour by other cartels? Can we flag potential offenders?

    Methods: Textual analysis, web crawling, sentiment analysis, R

    Supervisor: Evelina Gavrilova-Zoutman.

  • Detecting Corruption in the Oil-For-Food Program

    Detecting Corruption in the Oil-For-Food Program

    The Oil for Food Program (OFFP) was a relief effort orchestrated by the United Nations to help the people of Iraq after the Gulf War. It lasted from 1995 to 2003. Leaks from classified reports reveal that there has been rampant corruption, from the bank that handled the Iraq escrow account, to the trucking company that was supposed to handle the logistics of food transport. Even the then UN General Secretary Kofi Annan has been implicated in this corruption scandal. By looking at important events that influence the survival of the OFFP and stock prices of companies bidding for contracts, by virtue of insider trading, we can find an indirect proof for corruption.

    The methodology for this thesis is the same as in DellaVigna, S. and La Ferrara, E., 2010. Detecting illegal arms trade. American Economic Journal: Economic Policy, 2(4), pp.26-57.

    Supervisor: Evelina Gavrilova-Zoutman.

  • Detecting Illegal Arms Trade

    Detecting Illegal Arms Trade

    In an influential study from 2010 scholars have identified the likelihood that arms companies have been breaking embargo rules. The detection method hinges on insider trading as a result of war events that decrease or increase hostilities. The authors find a larger effect for companies from high corruption countries, where the cost of embargo violation is low. The authors also provide a methodology to detect individual culprits, who are likely to have sold arms in countries under an embargo.

    The data for this study spans 1990 to 2005. The question is: do these findings still hold in 2005 to 2020? Are the perpetrators different? For the whole data: Can we link perpetrators through their board composition? How connected are the boards of companies that violate embargoes?

    Additional questions: How does corporate governance influence illegal trading? Does the presence of an extensive network of offshore affiliate increase or decrease the stock price reaction to adverse events?

    Starting point: DellaVigna, S. and La Ferrara, E., 2010. Detecting illegal arms trade. American Economic Journal: Economic Policy, 2(4), pp.26-57.

    Supervisor: Evelina Gavrilova-Zoutman.

Data Science and Analytics

  • Text analysis and patents

    Text analysis and patents

    Patents are legal documents that are used by innovators to protect their innovations. The aim of the thesis is to use text analysis of different kinds to use the vast amount of information contained in the documents. This information can be potentially used to identify patenting trends, to understand the innovation process better, to improve our understanding of the importance of IPR etc. The information in the text can also be combined with Machine Learning techniques to predict, for example, the success of an application, the likelihood of an opposition or the likelihood of an infringement.

    Supervisor: Steffen Juranek.

  • Nonlinear econometrics

    Nonlinear econometrics

    A large portion of empirical research within economics and finance is based on linear models, of which the linear regression is by far the most prominent. Is this because we live in a linear world, or at least an approximately linear world, or is it the case that we implicitly close our eyes to important features i our data by not considering nonlinear methods on equal footing as traditional ones?

    Questions like this may take your master project in several directions, such as (listed from least to most statistical/mathematical maturity recommended to complete the project, all of them benefit from programming skills):

    1. To what degree is linear regression the main vehicle for measuring marginal effects of explanatory variables X to a response variable Y within economic research (within a certain field/ in Norway/ at NHH or otherwise suitably limited)? Why do researchers choose this method (convenience/interpretability of coefficients/easy presentation/theoretical foundations/...)?  Then, figure out to which degree such concerns can be addressed by a corresponding nonlinear model. A nice touch would be to re-do a recent linear study nonlinearly and see if there indeed are effects that were missed.

    2. A bit more technical version of the point above is to write a thesis that revolves around the systematical development of a tool in your programming language of choice (such as R or Python) that implements as many needs as possible of the linearly oriented researcher in a nonlinear framework, with pre-work consisting of providing a theoretical foundation, and as post-work perhaps testing your «product» by trying to «sell it» to an experienced researcher.

    3. Financial time series are typically nonlinear in the sense that the correlation coefficient does not, in general, give good descriptions of dependencies across time and space. This has naturally lead to the development of nonlinear methods to model financial processes. For example, the classical theory for portfolio allocation that Harry Markowitz introduced in the 1950s balances expected return (as measured by means) and risk (as measured by standard deviations and correlations) in order to provide the optimal distribution of wealth across different assets.

    The Markowitz portfolio theory is very simple and easy to implement. But, it explicitly assumes that dependence between assets is linear, so the decades following its introduction have seen many attempts to improve the Markowitz method by modelling dependencies nonlinearly. Many authors note, however, that it is actually quite hard to attain higher returns using modern methods compared to the classical approach. This project may contain a survey of modern portfolio selection methods (which will require the ability to read fairly technical research papers), and a discussion part where we try to answer the question whether beating the classical approach indeed is «hard», and if so, why?

    Supervisor: Håkon Otneim.

  • Sports Analytics

    Sports Analytics

    Due to its massive popularity and often large availability of data, sports present great opportunities for the application of analytics techniques. I have dedicated a great share of my research to topics related to sports analytics, including tournament scheduling, referee assignment, fairness, and ranking design. These are only examples from the broad range of topics in the agenda of sports analytics nowadays. There is a lot of literature about it. I would be open to discuss your specific interests and to provide you with references that could set the basis for a potentially fun and relevant master thesis.

    Supervisor: Mario Guajardo.

  • Using Textual Analysis to identify whether there is a gender gap in financial white collar crime

    Using Textual Analysis to identify whether there is a gender gap in financial white collar crime

    There is very little systematic evidence on the gender gap in crime. In particular, it is difficult to quantify it for financial white collar crime, as they do not find their way into police statistics and as there is little female presence on the top levels of companies.

    In this project, we can use textual analysis tools to collect data from the Securities and Exchange Commission (SEC) on insider trading and other frauds. We can match the name of the defendant to a gender, and quantify what is the difference between males and females.

    Then, we can correlate the gap, as well as the fraud itself, to past financial statements of involved companies in terms of gender representation in the board of the company and other indicators of company culture. Finally, we can compare the gap to other measures of female participation in the boardroom and determine whether white collar females seem more or less likely to commit crimes than white collar males. We can provide a partial answer to the question: Are companies going to become more responsible (do less criminal rule-breaking) if there are more females on the board?

    Methods: Textual analysis, web crawling, R

    Supervisor: Evelina Gavrilova-Zoutman.

  • What do companies hide? Geographic disclosures, tax haven presence and earnings statements

    What do companies hide? Geographic disclosures, tax haven presence and earnings statements

    Using textual analysis on earnings statements from the Securities and Exchange Commission EDGAR database, we can determine what do companies disclose about their operations in tax havens.

    On one hand, it is common knowledge that companies, who have affiliates in tax havens, do not actually conduct any real economic activity in these tax havens. They use these affiliates to avoid taxation through profit shifting. Through textual analysis, we can determine in what context are tax havens mentioned in financial statements and whether there is real activity associated with them. We can compare that to actual company statements and determine whether there is a difference between claims and reality.

    Do companies discriminate between tax havens and secrecy havens? Are they more likely to name specifically secrecy jurisdictions in the geographic disclosure, because they know that these countries would never reveal any information? Or, would the naming of such an affiliate be a bad signal for the presence of shady activities? Has this affiliate been named in press releases, if yes, with what sentiment?

    Methods: Textual analysis, web crawling, R

    Starting point: H Akamah, OK Hope, WB Thomas - Journal of International Business Studies, 2018

    Supervisor: Evelina Gavrilova-Zoutman.

  • Who blows the whistle on cartels?

    Who blows the whistle on cartels?

    By using webcrawling and textual analysis, you can gather data from the EU commission website on who reports the presence of market cartels. What is the reward that these whistleblowers get? Did they participate in the cartel before they reported it? Are these companies characterized by better corporate governance? Why did they blow the whistle? Could it be because of big shifts in governance structure, new boards, etc? What are sentiments on social networks on the mentions of the cartels and on the activity of the whistleblower?

    Are cartels more likely to be reported in certain industries? By using the revealed data on these cartels, how can we tailor a detection strategy that would detect similar behaviour by other cartels? Can we flag potential offenders?

    Methods: Textual analysis, web crawling, sentiment analysis, R

    Supervisor: Evelina Gavrilova-Zoutman.

  • Correcting witness reports through Machine Learning

    Correcting witness reports through Machine Learning

    It is well known that witness accounts on crime are often unreliable. The effects of stress or poor light on the victim can create a skewed perception of how the perpetrator looked. This, coupled with an unconscious discriminatory bias, leads to the victim reporting that the perpetrator was unknown or from the black racial minority. Often the witness report is based on an estimate – e.g. “ the perpetrator looked black, around 25 years”.

    Comparing actual arrests to reports, we can try to put a number on the amount of witness error. By training a machine learning model on a subsample, we can try to predict the gender, race and age of unobserved criminals. This can impact the computation of statistics on black/white, male/female and young/old crime gaps. What if blacks are actually responsible for a very small fraction of crimes? What if females are responsible for more crimes? This could lead us to rethink current racial profiling strategies in crime detection.

    Data: National Incident Based Reporting System, US

    Starting point: Imbens, G.W. and Lemieux, T., 2008. Regression discontinuity designs: A guide to practice. Journal of econometrics, 142(2), pp.615-635.
    Fryer Jr, R.G., 2016. An empirical analysis of racial differences in police use of force (No. w22399). National Bureau of Economic Research

    Supervisor: Evelina Gavrilova-Zoutman.

  • When the police is cheating

    When the police is cheating

    Police agencies get funding based on the amount of crime in their jurisdiction. This gives them the incentive to manipulate crime statistics. This can mean that for e.g. aggravated assaults are downgraded to simple assaults, or simple crimes are elevated to felonies. The police can add drug charges, in order to appear as being successful against organized crime. All this behaviour is hidden behind crime statistics.

    Through a combination of machine learning and local linear estimates we can try to determine how police agencies cheat. With the use of election outcomes as instrumental variables, we can try to find a causal effect. The findings of this project could have a strong impact on how crime figures are viewed. With corrected crime figures we can replicate previous analyses and determine whether well known policies actually impact crime or not.

    A separate question would be how do policing incentives influence police cheating in reporting statistics? E.g. federal grants that increase the police labor force vs. militarization donations that increase police capital. Both of these types of policies expect a crime decrease as a result and this is what we find in the literature, but is this finding warranted? Do we observe more crime downgrading after the policies than before?

    Data: National Incident Based Reporting System, US

    Supervisor: Evelina Gavrilova-Zoutman.

Energy Markets, Resource Management and Sustainability

Incentives, Contracts and Firm Behaviour

  • Economics of organization and management control

    Economics of organization and management control

    Including the use of bonus pay (case studies or across companies), how to measure performance, balanced scorecard (or alternative ways to use key performance indicators for management control purposes), drivers of profitability in an industry or a company, beyond budgeting, transfer pricing, and organizational boundaries.

    Supervisor: Iver Bragelien and Malin Arve.

  • Topics on Organizational Economics and Industrial Organization

    Topics on Organizational Economics and Industrial Organization

    Some examples from the past:

    • Om kontrakter og incentiver i ERP-implementering
    • En teoretisk analyse av innovasjonsinsentivene i Posten Norge AS
    • Digitalisering i Bygg- og anleggsnæringen
    • Kundelojalitetsprogrammer i det norske dagligvaremarkedet
    • Utviklingen av markedsstrukturen i den norske dagligvarebransjen
    • Lønnsomhetsforskjeller i større foretak (norsk møbelbransje)

    Supervisor: Trond E. Olsen and Malin Arve.

  • How did the US Tax Bill affect M&As?

    How did the US Tax Bill affect M&As?

    The 2017 TCJA may significantly change the pattern of corporate M&As. For instance, it has lifted the original use requirement for immediate depreciation of acquired assets. That makes it much more attractive to structure deals as asset purchases rather than stock acquisitions.

    Furthermore, net operating losses before and after January 1, 2018 are treated differently. An empirical analysis could use either Zephyr or SDC Platinum to explore early signs of these changes. 

    Supervisor: Maximilian Todtenhaupt.

  • Common ownership and tax avoidance

    Common ownership and tax avoidance

    Corporate tax planning is high on the agenda of both tax practitioners and policy makers. While some firms certainly try to avoid paying taxes, we know very little about how firms learn about different tax avoidance mechanisms. An important channel may be via common owners. Do firms engage in similar levels of tax avoidance if they are owned by the same investors? This project will investigate this question using balance sheet information and ownership details of large corporations.

    Supervisor: Maximilian Todtenhaupt.

  • Women and innovation

    Women and innovation

    It can be empirically observed that women are underrepresented in science and research. This is likely to have negative consequences for firm productivity and economic growth because talent is not used efficiently. 
    There are several ways by which a Master Thesis can contribute in identifying potential causes and remedies. Examples are the analysis of the examination procedure or an analysis of geographical patterns.

    Supervisor: Steffen Juranek.

  • Text analysis and patents

    Text analysis and patents

    Patents are legal documents that are used by innovators to protect their innovations. The aim of the thesis is to use text analysis of different kinds to use the vast amount of information contained in the documents. This information can be potentially used to identify patenting trends, to understand the innovation process better, to improve our understanding of the importance of IPR etc. The information in the text can also be combined with Machine Learning techniques to predict, for example, the success of an application, the likelihood of an opposition or the likelihood of an infringement.

    Supervisor: Steffen Juranek.

  • Taxation and Innovation

    Taxation and Innovation

    In a recent working paper Akcigit et al (2018) use US data to show that higher income tax rates reduce innovation. The aim of this thesis is to find out if a similar relationship between innovation and taxation exists for Europe. Students will use the universe of patent applications to the European patent office to find out i.) if countries with lower tax rates file for more patents and ii.) whether changes in the tax rate cause changes in patent applications. The outcome is particularly relevant for policy makers, since innovation is an important driver of economic growth. If higher tax rates indeed reduce innovation, this provides a strong incentive for countries not to increase their tax rate, or even to reduce it.

    Data: Patent applications in Europe

    Literature: Akcigit, U., Grigsby, J., Nicholas, T., & Stantcheva, S. (2018). Taxation and Innovation in the 20th Century. NBER Working Paper.

    Supervisor: Floris Zoutman and Steffen Juranek

  • The publicity effect of patent lawsuits

    The publicity effect of patent lawsuits

    Citations to patents matter because they indicate knowledge spillovers and indicate the importance of a technology. It is well-known that patents with more citations are more likely to be taken to court. However, it is less clear whether there is a publicity effect of lawsuits, i.e., whether lawsuits lead to more citations - potentially indicating that lawsuits enhance knowledge dispersion. The thesis should empricially analyze the latter effect.

    Supervisor: Steffen Juranek.

  • The stock market effect of hackers

    The stock market effect of hackers

    Software security became a more and more important topic in recent years. Hackers are a  threat to software producers around the world. This thesis aims to quantify the effect on the stock price of software producers after a hacking attack in a event study.

    Supervisor: Steffen Juranek.

  • Who blows the whistle on cartels?

    Who blows the whistle on cartels?

    By using webcrawling and textual analysis, you can gather data from the EU commission website on who reports the presence of market cartels. What is the reward that these whistleblowers get? Did they participate in the cartel before they reported it? Are these companies characterized by better corporate governance? Why did they blow the whistle? Could it be because of big shifts in governance structure, new boards, etc? What are sentiments on social networks on the mentions of the cartels and on the activity of the whistleblower?

    Are cartels more likely to be reported in certain industries? By using the revealed data on these cartels, how can we tailor a detection strategy that would detect similar behaviour by other cartels? Can we flag potential offenders?

    Methods: Textual analysis, web crawling, sentiment analysis, R

    Supervisor: Evelina Gavrilova-Zoutman.

  • Sivilrettslig inndragning av utbytte fra kriminelle handlinger

    Sivilrettslig inndragning av utbytte fra kriminelle handlinger

    JD nå jobber med forslag til lov om sivilrettslig inndragning (etter en utredning av prof. Jon Petter Rui). Tidligere er det blitt foreslått en nasjonal enhet som skal sikre fratagelse av verdier gjennom både sivilrettslige og strafferettslige verktøy. Disse to tiltakene ble utredet parallelt. Regjeringen sa nylig at dette hadde noen praktiske hindre, uten å gå nærmere inn på hva disse er. Er sivilrettslig inndragning et effektivt virkemiddel? Kan det bøte på det ofte rapporterte problemet om manglende kapasitet og/eller kompetanse i politidistriktet? Hva er de praktiske hindringene regjeringen refererer til? Hva er erfaringene i andre land som har brukt dette, hva er potensielle fallgruver og hva har hatt best effekt? Hvor stor er den preventive effekten av å ramme utbyttet/profitten fra den kriminelle virksomheten? Samarbeidsforum kan bidra som diskusjonspartner.

    In cooperation with Skatteetaten.

    Supervisor: Evelina Gavrilova-Zoutman.

  • Telenor-boten

    Telenor-boten

    Konkurransetilsynet gav i juni Telenor 788 mill. i bot. De mener Telenor har misbrukt sin markedsmakt ved å forhindre at et tredje mobilnett ble utbygd, og ilegger selskapet 788 millioner i overtredelsesgebyr. Telenor nekter for å ha gjort noe galt. Hva ligger bak denne boten? Er den korrekt?

    Krav: God bakgrunn i IO. Konkurranserett bra, men ikke nødvendig.

    Supervisor: Øystein Foros.

  • Telia-Get fusjonen

    Telia-Get fusjonen

    Gjøre en konkurranseanalyse av media/telemarkedet.

    Supervisor: Øystein Foros.

  • Konkurransen i mobilmarkedet

    Konkurransen i mobilmarkedet

    Sammenligne f.eks. Norge og Finland som har ulik markedsstruktur. Mye deskriptiv empiri for å bedre forstå markedet og forskjellen mellom landene.

    Supervisor: Øystein Foros.

  • Bokavtalen

    Bokavtalen

    Kulturminister Trine Skei Grande har nylig foreslått å utvide bokavtalen. Det vil gi en lengre fastprisperiode og høyere priser på opptil 65 prosent av bøkene. Den norske bokavtalen, som gir forleggerne rett til å sette en fast pris på nye bøker, er havnet i søkelyset til Efta-landenes overvåkningsorgan Esa. Nylig besvarte Nærings- og fiskeridepartementet (NFD), på vegne av den norske regjeringen, en rekke spørsmål fra Esa om forholdet mellom bokavtalen og EØS-avtalens eksplisitte forbud mot prissamarbeid i artikkel 53. Her kunne det være interessant å sammenligne med kommisjonens sak mot Amazon ift most-favored nation (MFN) klausuler.

    Supervisor: Øystein Foros.

  • Personalisert prising

    Personalisert prising

    Med utgangspunkt i prosjektet Moving towards the market of one? Competition with personalized pricing and endogenous mismatch costs, jobbe med applikasjoner? Kan også knyttet opp mot AI.

    Supervisor: Øystein Foros.

Shipping, Logistics and Operations Management

  • Shipping: How can CO2 emissions be priced in chartering contracts?

    Shipping: How can CO2 emissions be priced in chartering contracts?

    Cargo owners and charterers are the key to reducing emissions in global seaborne transportation. How can we revise contracts such that all stakeholders have an incentive to perform efficiently?

    Supervisor: Professor Roar Ådland

  • Shipping: Benchmarking charterer/owner performance in the chartering market

    Shipping: Benchmarking charterer/owner performance in the chartering market

    Using public fixture data, to analyze the extent to which company name affects market power in negotiations and freight rates. Requires knowledge of fixed-effects regression models.

    Supervisor: Professor Roar Ådland

  • Shipping: Economic optimization of underwater hull-cleaning intervals

    Shipping: Economic optimization of underwater hull-cleaning intervals

    Using fuel consumption and cost data from individual ships: How often should the company clean the hull of a ship to reduce fuel costs. Advanced – requires some knowledge of optimization, optimal stopping problems, semi-parametric regressions.

    Company: SKS Tankers.

    Supervisor: Professor Roar Ådland

  • Shipping: How can vessel performance data be used to optimize commercial operation?

    Shipping: How can vessel performance data be used to optimize commercial operation?

    Using vessel performance and open-source weather data – show how better fuel cost estimates affect the economics of ship operation. Strong knowledge of econometrics required.

    Company: SKS Tankers or Golden Ocean.

    Supervisor: Professor Roar Ådland

  • Shipping: The economics of IMO 2020

    Shipping: The economics of IMO 2020

    Using updated data on timecharter and/or voyage charter rates, investigate whether freight rates reacted to the new regulations implemented on January 1, 2020.

    Supervisor: Roar Ådland

  • Shipping: Vessel speed analysis using AIS data

    Shipping: Vessel speed analysis using AIS data

    Using AIS data on ship positions, investigating how vessel speeds react to changes in fuel prices and spot freight rates in the short- and long run.
    Requires familiarity with co-integration tests and econometrics.

    Supervisor: Professor Roar Ådland

  • Shipping: Network analysis in the international oil market

    Shipping: Network analysis in the international oil market

    Using micro-data on the buyer and seller of crude oil cargoes, to investigate the structure of commercial relationships in the global crude oil market and how it changed with the drop in the oil price in 2014. Part of this is related to risk management – do agents have a portfolio of counterparts or do they rely on one source?

    Supervisor: Professor Roar Ådland

  • Shipping & Finance: Machine learning models for FFA trading

    Shipping & Finance: Machine learning models for FFA trading

    Using spatial AIS data for ship positions, open-source weather and macro data – can you develop a machine learning model to generate profitable trading signals? Requires knowledge of Python, implementation of machine learning models.

    Supervisor: Roar Ådland and Haiying Jia

  • Shipping & Finance: How do risk capital and risk limits affect the chartering policy of a ship operator?

    Shipping & Finance: How do risk capital and risk limits affect the chartering policy of a ship operator?

    Using: freight rates timeseries, optimal portfolio theory.

    Company: Western Bulk.

    Supervisor: Professor Roar Ådland

  • Shipping & Finance: Inferring short-term market direction from intraday FFA Data

    Shipping & Finance: Inferring short-term market direction from intraday FFA Data

    Using intraday bid/offers spreads – is it possible to use pattern recognition or technical analysis to daytrade forward freight agreements? Knowledge of Python and machine learning is useful. Can possibly be done in cooperation with Clarksons.

    Supervisor: Professor Roar Ådland

  • Alternative shipping capital sources

    Alternative shipping capital sources

    The development of Chinese shipping trusts as alternative shipping finance sources.

    Supervisor: Haiying Jia

  • Blockchain and smart contracts in shipping and supply chain management

    Blockchain and smart contracts in shipping and supply chain management

    How will blockchain change the relationships among shipping players? Which areas do blockchain provide the most value for the shipping ecosystem? May be survey based. Incentive mechanism in a blockchain-based smart contract in shipping - how to incentivize players to participate?

    Supervisor: Haiying Jia

  • Risk modelling in ship finance and marine insurance

    Risk modelling in ship finance and marine insurance

    How to aggregate risk exposure in the content of credit risk, geographical exposure, and event risks etc.? How to optimally transfer risks in reinsurance? How to evaluate financial returns with regards to low probability - high severity risks?

    Supervisor: Haiying Jia

  • Sustainable ship finance

    Sustainable ship finance

    How ship finance helps to achieve green shipping, using Poseidon Principles as a case study.

    Supervisor: Haiying Jia

  • Blå Bybane (Blue Light Rail)

    Blå Bybane (Blue Light Rail)

    Bergen has built its first light rail line from the city centre to the airport. Other lines are being discussed politically. This thesis covers a related idea, which has been called “Blue Light Rail”, though it is not a rail line at all, but an idea of having scheduled vessels (high speed boats) playing the same role as light rail, and utilizing the fact that Bergen is a coastal city with complicated topology.

    There are several places with a reasonably large transportation need that are more easily reached from Bergen city centre by boat than by car or bus (or rail). The most obvious ones are Askøy, Knarvik (which already have connections) and Ågotnes (Sotra), which has not. It is also possible to imagine routes within the area around the city centre itself.

    A suggestion has been to connect Laksevåg, Dokken, Hegreneset and Sandviken (if you don’t know Bergen, ask Mr Google). This could facilitate a fascinating development of central Bergen by providing clean and efficient transportation to new developments. Typically these vessels will have electrical engines.

    Bergen Chamber of Commerce (Næringsrådet) has established a committee looking into the Blue Light Rail, and this thesis will interact with that committee. The thesis is also placed within a project on autonomous vessels (though autonomy comes later) funded by the Norwegian Research Council, attached to NHH’s Centre for Shipping and Logistics.

    The thesis can take many forms related to such as scheduling relative to other means of public transportation (possibly with a special emphasis on how delays can be handled), the interaction with tourism (it is quite attractive for some groups of tourists to get a trip on the fjord at ordinary, non-touristic prices), and an analysis of the environmental advantages of such a service. But feel free to suggest other perspectives on the problem, as there are others.

    The thesis can fit into many different profiles, particularly BAN, ENE and BUS, depending on the profile of the students.

    NHH contact: Stein W. Wallace

    External contacts: This depends on the profile of the thesis. Possibilities are:

    • Bergen Chamber of Commerce (by its leader Atle Kvamme) and the Chamber’s Blue Light Rail Committee. This covers almost any profile you could possibly choose.
    • CEO Johnny Breivik, Port of Bergen
    • Member of the municipal government (with responsibility for city development) Anna Elisa Tryti and her advisor Tarje Wanvik.
    • Roger Harkestad at TIDE (bus company)
    • Lars Jacob Engelsen at Norled (boat and ferry company)

     

  • Small city logistics

    Small city logistics

    Urban population growth is driving an increase in the amount of freight that goes into and out of cities. That growth poses an increasing challenge to freight transportation in smaller compact cities with difficult topology, which is typical for most Norwegian cities and numerous cities abroad. This transportation challenge is exacerbated by phenomena such as an increase in internet trade, the demand for fast delivery, and a reduction in the ownership of private cars in the city centre which could be used for shopping. The result is an increase in the total volume of freight, and more critically, in the total number of deliveries, normally managed by a large variety of transportation companies. Unless planned for and regulated, a consequence might be increased traffic, with enhanced energy consumption, that competes for available space and may affect living conditions for a growing urban population.

    This project will study small city logistics, with a focus on Bergen, to find the options available for the authorities, as well as business models for a better city logistics setup. Will be done in cooperation with the City of Bergen and Bergen Chamber of Commerce. The project can be qualitative as well as quantitative.

    Supervisor: Stein W. Wallace.

  • How will autonomous vessels change the operations in the shipping industry (including deep-sea shipping, short-sea shipping and local waterway transport)?

    How will autonomous vessels change the operations in the shipping industry (including deep-sea shipping, short-sea shipping and local waterway transport)?

    Background

    If you are a hunter from the Stone Age and one day you are facing an offer to replace you wooden stick with a brand new shotgun, will you still use your new weapon just as a harder stick made of steel to kill your prey, or use it in a better way? Similar challenges are now faced by the shipping companies due to the forthcoming technological evolution, namely the autonomous ship. Obviously, an autonomous ship with no crews on board can significantly reduce a shipping company’s crew cost.

    However, just like the increased hardness of a shotgun in the hunter example, the reduction of crew cost might just be a tiny benefit of the autonomy of our ships. Besides the lower crew cost, what are the fundamental advantages of an autonomous vessel comparing to the conventional manned ship?

    Greater potentials are expected by better utilizing these advantages with innovative ideas in the daily operation of the vessels, such as higher frequency of ferry in the night time, flexible hub location for waterway taxi and multi-functioned vessels with different remote control teams. The world’s first commercial autonomous vessel (Yara Birkeland) will be soon launched in Norway in the end of 2018. And it is a great opportunity for the students here to also take the leading position in the research of the autonomous vessel.

    Potential Reference

    Toth, P., Vigo, D. (2002). The Vehicle Routing Problem. Philadelphia: SIAM Publications Stopford, M. (2009). Maritime Economics. London: Routledge.

    Contact: Yewen Gu

  • Using autonomous vehicles to improve our emergency services

    Using autonomous vehicles to improve our emergency services

    The aim of emergency medical services (EMS) is to provide timely assistance to emergencies in order to save lives. Within this service, quality and capacity have sometimes deteriorated because staffing is not satisfactory and because the organization and directives are not clear. My interest is to work on the use of autonomous vessels to help ameliorate the burden that EMS staffing represents in the case of boat ambulances, and to improve the logistics planning of the system.

    The aim is to analyze the use of autonomous vessels to improve response times and coverage. For example, by combining autonomous vessels with geographic information systems, one may use real time information of potential patients to improve the deployment of the resources. In particular my interest is to explore the following key research topic: designing algorithms with predictive capabilities that can be included in real time systems and capable of managing a continuous feed of data points coming from users’ cell phones and other sources.

    Supervisor: Julio C. Goez.

  • The Operational and Economic Impact of Autonomous Ship Application, Comparing to the Traditional Manned Vessels. Wartsila, Fjellstrand, NYK

    The Operational and Economic Impact of Autonomous Ship Application, Comparing to the Traditional Manned Vessels. Wartsila, Fjellstrand, NYK

    Supervisor: Yewen Gu.

  • Repositioning of Empty Vessels in the Dry Bulk Shipping Market

    Repositioning of Empty Vessels in the Dry Bulk Shipping Market

    Aim: find key drivers for decision-making process of repositioning empty vessels - current market conditions, sentiment - repeating patterns, etc.

    Supervisor: Vít Procházka.

  • Emission Abatement Technology for a Shipping Company - is the Uncertainty of Fuel Prices Important?

    Emission Abatement Technology for a Shipping Company - is the Uncertainty of Fuel Prices Important?

    Is the uncertainty of fuel prices important to be considered when a shipping company selects its emission abatement technology for the compliance of ECA regulation.

    Supervisor: Yewen Gu.

  • Operations Research Applications in Tine

    Operations Research Applications in Tine

    Dairy farmers with combined milk and beef production face complex decisions regarding optimum milk yield, slaughter age for bulls, calving age for heifers, disposal of farm land etc. The aim for this topic is to explore how operations research may help farmers improve their decision making process to increase their profitability. In this project the students will interact with TINE, Norway's largest producer, distributor and exporter of dairy products with 11,400 members (owners) and 9,000 cooperative farms.

    Supervisors: Mario Guajardo and Julio C. Goez .

  • Logistics/sharing economy: Analytics for car-sharing models

    Logistics/sharing economy: Analytics for car-sharing models

    Car-sharing provides short-term vehicle access to a group of user members who share the use of a vehicle fleet owned by a car-sharing organization that maintains, manages, and insures the vehicles. An example of this model in Bergen is bildeleringer. Managing the fleet involves decisions such as the size of the fleet, how to position and reposition the vehicles, maintenance schedules, and pricing approaches. Strong background on analytics required

    Supervisors: Julio C. Goez and Mario Guajadaro.

  • Optimization of requirements of cloud computing resources

    Optimization of requirements of cloud computing resources

    The providers of online applications usually need to find the deployment of minimum cost for running it in the cloud. For the deployment, the planner on the application side must consider renting resources from cloud providers. However, there is a service level constraint that must be satisfied to ensure the quality of the service.

    Supervisor: Julio C. Goez.