Browsing by study line "Yleinen opintosuunta"

Taxation is a critical tool for development, as well-designed tax systems can generate greater revenues to fund public goods and investments that drive productivity. However, lower-income countries often raise only a small percentage of their gross domestic product (GDP) in taxes compared to higher-income countries. The gap in revenue is particularly striking for property taxes. This thesis begins with a literature review on taxation in developing countries and property taxation in general, emphasizing the challenges of property taxation in developing countries. It then provides relevant background information on Kenya, discussing inequality, general taxation, and property taxation. Using survey data from the Kenya Integrated Household Budget Survey 2015-2016, the thesis then investigates the distribution of property value and income among Kenyan households. It then explores the potential revenue and distributional effects of two property taxation models while also considering potential revenue loss from taxing households with insufficient income or savings. The trends between household income and property value indicate that higher annual incomes correlate with higher property values and vice versa. However, low-income individuals also owned higher-value properties, leading to liquidity problems. A 2% linear property tax rate model revealed a heavy tax burden on lower-income households and an overall revenue potential of 2.38% of total survey income or 0.94% of 2016 GDP. The study observed significant decreases in revenue potential when adjusting revenue to account for the ability to pay. This research also modeled a property tax rate payable only upon reaching a specific income threshold to address liquidity problems. In this model, the tax burden shifted to higher-income households with an overall potential tax revenue of 1.91% of total survey income or 0.76% of GDP. In conclusion, the observed trends, such as the high prevalence of inability to pay, relatively low revenue potential, and the administrative effort required for property taxation, suggest that reforming property taxation may not be the most practical approach for increasing revenue in Kenya.

Machine Learning (ML) has experienced significant growth, fuelled by the surge in big data. Organizations leverage ML techniques to take advantage of the data. So far, the focus has predominantly been on increasing the value by developing ML algorithms. Another option would be to optimize resource consumption to reach cost optimality. This thesis contributes to cost optimality by identifying and testing frameworks that enable organizations to make informed decisions on cost-effective cloud infrastructure while designing and developing ML workflows. The two frameworks we introduce to model Cost Optimality are: "Cost Optimal Query Processing in the Cloud" for data pipelines and "PALEO" for ML model training pipelines. The latter focuses on estimating the training time needed to train a Neural Net, while the first one is more generic in assessing cost-optimal cloud setup for query processing. Through the literature review, we show that it is critical to consider both the data and ML training aspects when designing a cost-optimal ML workflow. Our results indicate that the frameworks provide accurate estimates about cost-optimal hardware configuration in the cloud for ML workflow. There are deviations when we dive into the details: our chosen version of the Cost Optimal Model does not consider the impact of larger memory. Also, the frameworks do not provide accurate execution time estimates: PALEO estimates our accelerated EC2 instance to execute the training workload with half of the time it took. However, the purpose of the study was not to provide accurate execution or cost estimates, but we aimed to see if the frameworks estimate the cost-optimal cloud infrastructure setup among the five EC2 instances that we chose to execute our three different workloads.

The importance of protecting sensitive data from information breaches has increased in recent years due to companies and other institutions gathering massive datasets about their customers, including personally identifiable information. Differential privacy is one of the state-of-the-art methods for providing provable privacy to these datasets, protecting them from adversarial attacks. This thesis focuses on studying existing differentially private random forest (DPRF) algorithms, comparing them, and constructing a version of the DPRF algorithm based on these algorithms. Twelve articles from the late 2000s to 2022, each implementing a version of the DPRF algorithm, are included in the review of previous work. The created algorithm, called DPRF_thesis , uses a privatized median as a method for splitting internal nodes of the decision trees. The class counts of the leaf-nodes are made with the exponential mechanism. Tests on the DPRF_thesis algorithm were run on three binary classification UCI datasets, and the accuracy results were mostly comparable with the two existing DPRF algorithms DPRF_thesis was compared to. ACM Computing Classification System (CCS): Computing methodologies → Machine learning → Machine learning approaches → Classification and regression trees Security and privacy → Database and storage security → Data anonymization and sanitization

Carbon pricing is gaining momentum across developing countries such as the Philippines. Along with its strong commitment to increase the shares of renewable energy sources and excise taxes on fuels already in place, the government passed the House Bill No. 4939 aiming to tax households Php 1 for every kilogram of $CO_2$ emission per kilowatt hour (kWh) of power consumption. This raises varying opinions in the public if carbon tax is necessary compared to the amount of emissions of the country as a whole, amounting only to .48\% of total global emissions. Existing literature yields diverse results on the progressiveness of carbon tax in terms of its distributional effects. Most developed countries with active carbon taxation prove it to be progressive across different income groups. However, in developing countries such as the Philippines, studies show varying outcomes -- some argue that carbon tax would be regressive and would hurt the most vulnerable populations. Others also say that it would most affect the middle class, while few conclude that it can progressive in some developing countries. The Almost Ideal Demand System (AIDS) is used in this thesis to determine the change in household consumption across different income groups in the Philippines. The results show a regressive reaction to the increase in the price of electricity. The simulation resulted to a nearly similar pattern in the overall changes in consumption by varying degrees. The changes are most substantial in the poorest and richest households, meanwhile slight changes in the middle groups. However, by taking the income and cross-price elasticity of demand, the commodities are determined to be whether a necessity or luxury good and a complement or substitute good. The maximizing household then decides from which commodities to cut back in order to compensate to the carbon tax on electricity. The lowest income has the most percentage increase in the consumption of electricity, compensating from the necessities such as food. The middle groups follow by less percentage increase. Meanwhile, the highest income decreases its budget share for electricity, making way to increase on its other more important expenses. The elasticity of demand for carbon-intensive goods plays a role in determining the distributional effects of a carbon tax. In this thesis, the Php 1 carbon tax would affect the lowest income household in substantial amount compared to their actual energy consumption. This concerns policy makers of the social equity, and fairness of carbon tax. The House Bill No. 4939, when approved, takes the first step to adapting carbon tax in the Philippines.

Kartellit ovat yksi vakavimmista keinoista rajoittaa markkinoiden vapaata kilpailua. Tämän takia on tärkeää, että kilpailupolitiikalla pyritään torjumaan niitä mahdollisimman tehokkaasti. Tänä päivänä keskeinen kilpailupolitiikan väline kartellien torjumiseksi on leniency- eli ilmiantojärjestelmä. Ilmiantojärjestelmällä luodaan kartellin muodostaneille yrityksille kannustin ilmiantaa kartelli kilpailuviranomaiselle lupaamalla ilmiantavalle yritykselle esimerkiksi lievennys kartellista kiinnijäämisen myötä aiheutuvaan seuraamusmaksuun. Tässä tutkielmassa tarkastellaan, miten ilmiantojärjestelmä kartelleja torjuvana kilpailupolitiikan välineenä vaikuttaa yritysten kartellikäyttäytymiseen. Tutkielma on kirjallisuuskatsaus, jossa syvennytään aiempaan taloustieteelliseen kirjallisuuteen ilmiantojärjestelmistä. Ilmiantojärjestelmän vaikutusten yritysten kartelliin ryhtymisen ja sen ilmiantamisen kannustimiin tarkastelu aloitetaan käsittelemällä yksityiskohtaisesti Mottan ja Polon (2003) malli ilmiantojärjestelmästä. Tämän jälkeen perehdytään ilmiantojärjestelmiä käsittelevään kokeellisen taloustieteen empiriaan käymällä läpi Apesteguian, Dufwenbergin ja Seltenin (2007) sekä Hinloopenin ja Soeteventin (2008) tutkimukset. Kokeellisen taloustieteen empiriaa hyödyntäen voidaan vertailla erilaisia ilmiantojärjestelmiä sekä tarkastella, miten teoreettiset mallit mahdollisesti toimivat käytännössä. Huomataan, etteivät teoreettisten mallien ennusteet ja empiiriset tulokset ilmiantojärjestelmän toimivuudesta ja tehokkuudesta välttämättä ole yhteneväisiä. Tarkastelun perusteella havaitaan lisäksi, että ilmiantojärjestelmällä voidaan onnistua sekä saamaan kartellin jäsenet tekemään ilmianto että estämään kartellin muodostaminen kokonaan. Toisaalta on myös mahdollista, että ilmiantojärjestelmä kannustaa kartelliin ryhtymiseen, mikäli esimerkiksi ilmiantamisen seurauksena on mahdollista saada positiivinen palkkio. Ilmiantojärjestelmän ominaisuuksilla, kuten seuraamusmaksun alennuksen suuruudella, onkin keskeinen vaikutus siihen, miten ilmiantojärjestelmä vaikuttaa yritysten kartellikäyttäytymiseen. Tutkielman lopussa syvennytään pohtimaan keskeisiä ilmiantojärjestelmän ominaisuuksia. Yhteenvetona optimaalisen ilmiantojärjestelmän kiteytetään olevan kompromissi sen kartelleja torjuvan ja stabiloivan vaikutuksen välillä.

The thesis aims to analyse the individual subsidy effects on electric vehicle uptake by a unique case study of ten countries. The sample countries analysed in the thesis are Norway, Finland, Denmark, Sweden, China, France, Germany, Italy, Spain, and the United Kingdom. Together these countries represent most of the electric vehicle market share in the world and are thus potent for analysing different subsidy effect. The electric vehicle market has grown considerably in the past decade, and the markets’ role in reaching climate objectives set out by governments and other institutions is pivotal. The role of government initiatives in shaping the electric vehicle market is reviewed and the standard characteristics are discussed. The literature review outlines the importance of purchase subsidies, infrastructure subsidies and governmental institutions in increasing the electric vehicle share globally. Purchase subsidies that are designed to decrease the upfront costs of purchasing a vehicle are agreed to be most important in increasing uptake. Charging infrastructure subsidies are pivotal for increasing the number of electric vehicles in any country. However, the role of subsidising charging stations is less apparent in increasing electric vehicle sales shares. Researchers mostly identify charging infrastructure provisioning as a support function for uptake, but some evidence exists for direct influence on increased sales shares. Government institutions are paramount in increasing the market share for electric vehicles. Governments can design effective long-term policies that account for important societal factors, which private incentive designs might not consider. The empirical strategy set out in the thesis utilises a panel dataset from the sample countries to estimate the individual subsidy effects on electric vehicle uptake. Firstly, the dataset is analysed by a pooled ordinary least squares method to identify the relationship of purchase, ownership tax, infrastructure subsidies and other post-purchase subsidies on electric vehicle sales share. Secondly, a difference-in-differences estimation is conducted to analyse more predictive outcomes of ownership tax- and other subsidies on sales share. Finally, a fixed effects model is utilised to account for unobserved time-invariant heterogeneity. The results are limited by a relatively small sample size due to data availability, and the consequent limitations in the empirical models. The results suggest that the role of post-purchase subsidies should be better understood for accurate policy designs, especially when the market is past its infancy phase. Finally, suggestions for future research are provided.

Tämän tutkimuksen tarkoituksena on näyttää, miten koronapandemia vaikutti asuntomarkkinoihin vuosien 2019–2021 aikana. Asuntomarkkinoilla tapahtuneita muutoksia tarkastellaan sekä globaalisti, että paikallisesti Suomen näkökulmasta. Tutkimuksessa arvioidaan miten eri mekanismien kautta koronapandemian vaikutukset levisivät asuntomarkkinoille ja tätä kautta kotitalouksiin. Aiemmat tutkimukset ovat osoittaneet koronapandemian aiheuttaneen hyvin moniulotteisia vaikutuksia asuntomarkkinoihin. Tämä tutkimus keskittyy käsittelemään näitä vaikutuksia pääasiassa asuntorahoittamisen näkökulmasta. Tämän lisäksi koronapandemian vaikutuksia tarkastellaan asuntojen hintojen, etätyön, muuttoliikkeen sekä taloudellisen epätasa-arvon tasolla. Tutkimuksen teoreettinen osuus pohjautuu elinkaarimalliin, jota sovelletaan asuntomarkkinoille. Tutkimus toteutettiin hyödyntämällä laajasti kirjallisuutta liittyen koronapandemian vaikutuksiin Yhdysvaltojen asuntomarkkinoilla. Kansainvälistä tutkimusta täydennettiin Suomen Tilastokeskuksen datalla, sekä Helsingin Seudun Osuuspankin tarjoamalla asuntorahoituksen aineistolla. Aineisto kerättiin yhteystyössä Helsingin Seudun Osuuspankin edustajien kanssa niin, että yksittäisen henkilön tunnistaminen aineiston perusteella oli mahdotonta. Asuntorahoituksen laajasta aineistosta oli mahdollista nähdä kattavasti päivämäärän tarkkuudella Suomen pääkaupunkiseudun asuntomarkkinan yleinen kehitys sekä trendit. Aineisto tarjosi tiedon liittyen lainan kokoon, takaisinmaksuaikaan sekä asunnon tyyppiin. Tilastokeskuksen data-aineistoja käytettiin vahvistamaan tuloksia ja tuomaan esille sellaista tietoa Suomen asuntomarkkinoista, johon asuntorahoituksen aineisto ei vastannut. Tutkimusmenetelmänä hyödynnettiin aikasarja-analyysiä, jota sovellettiin tutkimuksen kaikkiin aineistoihin. Analyysin tarkoituksena oli näyttää miten asuntomarkkinat ovat Suomessa kehittyneet ennen koronapandemiaa ja sen aikana. Tutkimuksen aikajanaksi valikoituivat vuodet 2016–2021. Analyysin tuloksia verrattiin kansainvälisiin tutkimuksiin ja pyrittiin löytämään yhteisiä koronapandemian laukaisemia mekanismeja, jotka vaikuttivat asuntomarkkinoihin. Tutkimuksen tulokset Suomen asuntomarkkinasta ovat pääasiassa linjassa globaalin asuntomarkkinan kanssa, mutta myös eroavaisuuksia pystyttiin löytämään. Yhteistä Suomalaisessa ja globaalissa asuntomarkkinassa on asuntojen hintojen kasvu pandemian aikana. Mekanismit hintojen nousussa kuitenkin erosivat sillä Yhdysvalloissa kasvua tuki halventuneet rahoitusolosuhteet, kun taas Suomessa voitiin nähdä etenkin ensiasunnon ostajien kasvaneen kysynnän vaikuttaneen positiivisesti hintoihin. Lisäksi Suomessa hintoja tuki kasvanut kysyntä asuntolainoissa, joiden määrät sekä laina-ajat kasvoivat koronapandemian aikana. Keskeinen ero hintojen nousussa liittyi alueelliseen kehitykseen. Suomen pääkaupunkiseudulla hinnat kasvoivat eniten aivan Helsingin keskustassa, kun taas Yhdysvalloissa hintojen kasvu esikaupunkialueilla oli keskustoja nopeampaa. Keskeisenä tuloksena tutkimuksesta löytyi lisäksi se, että Suomessa sekä Yhdysvalloissa pandemia kiihdytti varallisuuseroja kasvattaen näin taloudellista epätasa-arvoa asuntovelallisten kotitalouksien välillä. Asuntomarkkinoilla tapahtuneet muutokset koronapandemian aikana olivat hyvin laajat ja koskettivat niin matala- kuin korkeatuloisia asuntovelallisia. Keskeiset mekanismit liittyivät makrovakaus– ja rahapolitiikan tekemiin muutoksiin, jotka tukivat asuntomarkkinoiden kasvua etenkin Yhdysvalloissa. Suomessa pandemian aiheuttamat vaikutukset eivät näkyneet yhtä vahvoina kuin Yhdysvalloissa. Koronapandemia kuitenkin muokkasi merkittävästi asuntomarkkinoita maailmanlaajuisesti. Pandemian vaikutukset tulevat näkymään etenkin kotitalouksien velkaantuneisuudessa sekä taloudellisessa tilanteessa vielä pitkän aikaa.

The first wave of basic income experiments took place in North America between the 1960s and 1980s, and the second wave started in the early 2000s and is still ongoing globally. Experiments are used to gain knowledge about effects that basic income has on labour supply, poverty, and welfare. This thesis is a literature review that utilises labour supply theory to examine labour supply effects that are observed when experimenting basic income and negative income tax policies with randomised controlled trials. The purpose is to find possible causes of differing results with the help of labour supply theory and by carefully studying the characteristics of the experiments. The topic is relevant as experimental results can be used when making decisions about future welfare reforms. Previous research on basic income has been done by Banerjee, Niehaus, and Suri (2019), Ghatak and Maniquet (2019), and Hoynes and Rothstein (2019), who find that the implementation of a basic income reduces labour supply more in advanced countries than they do in developing countries. Robins (1985) and Hum and Simpson (1993) find that negative income tax experiments resulted in labour supply reductions. On the other hand, other authors find that a basic income decreases labour supply only for those who are generally supposed to work less: the elderly, those with disabilities or illnesses, mothers of young children, and children (De Paz-Báñez, Asensio-Coto, Sánchez-López, & Aceytuno, 2020). I conclude that a universal basic income redirects labour supply in developing countries from wage labour to self-employment. However, exact increases or decreases in total labour supply are unknown. Negative income taxes decrease labour supply by few weeks of full-time employment per year. Although there does not seem to be a clear causal effect between the variation of experimental design and the results reported by the 21st century experiments, factors such as the structure of the research sample, geographical area, type of poverty trap, and size of the increase in living standards created by basic income may affect the beneficiaries’ labour supply.

Modeling customer engagement assists a business in identifying the high risk and high potential customers. A way to define high risk and high potential customers in a Software-as-a-Service (SaaS) business is to define them as customers with high potential to churn or upgrade. Identifying the high risk and high potential customers in time can help the business retain and grow revenue. This thesis uses churn and upgrade prediction classifiers to define a customer engagement score for a SaaS business. The classifiers used and compared in the research were logistic regression, random forest and XGBoost. The classifiers were trained using data from the case-company containing customer data such as user count and feature usage. To tackle class imbalance, the models were also trained with oversampled training data. The hyperparameters of each classifier were optimised using grid search. After training the models, performance of the classifiers on a test data was evaluated. In the end, the XGBoost classifiers outperformed the other classifiers in churn prediction. In predicting customer upgrades, the results were more mixed. Feature importances were also calculated, and the results showed that the importances differ for churn and upgrade prediction.

This thesis regards non-Lambertian surfaces and their challenges, solutions and study in computer vision. The physical theory for understanding the phenomenon is built first, using the Lambertian reflectance model, which defines Lambertian surfaces as ideally diffuse surfaces, whose luminance is isotropic and the luminous intensity obeys Lambert's cosine law. From these two assumptions, non-Lambertian surfaces violate at least the cosine law and are consequently specularly reflecting surfaces, whose perceived brightness is dependent from the viewpoint. Thus non-Lambertian surfaces violate also brightness and colour constancies, which assume that the brightness and colour of same real-world points stays constant across images. These assumptions are used, for example, in tracking and feature matching and thus non-Lambertian surfaces pose complications for object reconstruction and navigation among other tasks in the field of computer vision. After formulating the theoretical foundation of necessary physics and a more general reflectance model called the bi-directional reflectance distribution function, a comprehensive literature review into significant studies regarding non-Lambertian surfaces is conducted. The primary topics of the survey include photometric stereo and navigation systems, while considering other potential fields, such as fusion methods and illumination invariance. The goal of the survey is to formulate a detailed and in-depth answer to what methods can be used to solve the challenges posed by non-Lambertian surfaces, what are these methods' strengths and weaknesses, what are the used datasets and what remains to be answered by further research. After the survey, a dataset is collected and presented, and an outline of another dataset to be published in an upcoming paper is presented. Then a general discussion about the survey and the study is undertaken and conclusions along with proposed future steps are introduced.