The dominant role of credit ratings, along with the failure of important FIs, exacerbated the 2008 crisis and caused further damage to European economies, which highlighted the need for effective regulation to prevent a reoccurrence. This thesis investigates the effect of EU and US recent regulatory reforms of the rating industry on the quality of credit ratings of financial institutions (FIs), as well as the impact of the new EU financial regulatory initiatives on the performance of FIs. The first empirical Chapter focuses on the EU reforms of credit rating agencies (CRAs) and provides evidence supporting the presence of a conservative rating bias in the post regulatory period, as increased scrutiny, fines and liability increase the cost of over rating. CRAs exhibit an unwarranted decrease in EU FI ratings, evidenced by an increase in false warning and a fall in the informativeness of FI rating downgrades in the post regulatory period. A subsequent rise in stock market responses to rating upgrades is consistent with CRAs expending greater effort to ensure they are justified. The second empirical Chapter focuses on the US reforms of CRAs and reports no significant impact on FI ratings, rather each CRA has responded differently to the passage of the US Dodd-Frank Act (DFA). There is, however, a significant reduction in stock market reactions to FI credit rating signals, consistent with diminishing reliance on credit ratings by market participants in the US. The third empirical Chapter builds and estimates a dynamic model of FI behaviour using discrete choice dynamic programming (DCDP). The model is used to simulate and examine the impact of regulations, including EU reforms of CRAs, capital adequacy regulation (Basel III), and the bail-in regime, on FIs' behaviour in the real economy. The results show that the shift to increasingly conservative rating behaviour triggered by the CRA reforms has caused FIs to respond by manipulating their capital ratios and to reduce lending activities. The results also show that more stringent capital requirements stimulate FIs to hold more capital, reduce lending and reveal a positive influence in reducing bank insolvency rates, particularly during the crisis period. The introduction of a bail-in regime reveals similar results, but crucially stimulates the adoption of a stable equilibrium (unlike Basel III). This thesis highlights drawbacks with the current regulatory reforms of the EU and US FI rating industries and suggests potential solutions. The thesis also informs the policy debate surrounding the best way to regulate both CRAs and FIs and ensure that there is not a reoccurrence of the problems present in the 2008 financial crisis.

This thesis studies different aspects of firm decisions by using a dynamic model. I estimate a dynamic model of the firm based on the trade-off theory of capital structure that endogenizes investment, leverage, and payout decisions. For the estimation of the model I use Efficient Method of Moments (EMM), which allows me to recover the structural parameters that best replicate the characteristics of the data. I start analyzing the question of whether target leverage varies over time. While this is a central issue in finance, there is no consensus in the literature on this point. I propose an explanation that reconciles some of the seemingly contradictory empirical evidence. The dynamic model generates a target leverage that changes over time and consistently reproduces the results of Lemmon, Roberts, and Zender (2008). These findings suggest that the time-varying target leverage assumption of the big bulk of the previous literature is not incompatible with the evidence presented by Lemmon, Roberts, and Zender (2008). Then I study how corporate income tax payments vary with the corporate income tax rate. The dynamic model produces a bell-shaped relationship between tax revenue and the tax rate that is consistent with the notion of the Laffer curve. The dynamic model generates the maximum tax revenue for a tax rate between 36% and 41%. Finally, I investigate the impact of financial constraints on investment decisions by firms. Model results show that investment-cash flow sensitivity is higher for less financially constrained firms. This result is consistent with Kaplan and Zingales (1997). The dynamic model also rationalizes why large and mature firms have a positive and significant investment-cash flow sensitivity.

The first chapter of this dissertation argues that privatization and other policy changes in China during the past decade had a direct effect on input prices. This is in contrast to most previous work, which instead emphasized the effect of privatization on productivity. I illustrate the importance of taking such differences into account by estimating the parameters of a static firm competition model using the Chinese Industrial Enterprises Database and calculating a measure of total factor productivity for each firm out of those estimates. The results of my analysis indicate that the percentage difference in productivity between private and state-owned firms may be overestimated by as much as 135 percentage points if the difference in input prices is not properly addressed. The second chapter of my dissertation establishes empirical support for such a difference in input prices by constructing a dynamic structural model of privatization, firm heterogeneity and industry evolution, and estimating its parameters using the Chinese Industrial Enterprises Database. My estimates of the model confirm many well documented institutional features about China's reform, including "grasp the large and let go of the small" policy, easy access to credit for state-owned enterprises (SOEs), and selection for privatization according to firms' probabilities of success. In addition, the estimated structural model provides the basic tool for policy simulations and enables us to see the effects of hypothetical policies, for example “grasp the small and let go of the large” and “removing SOEs’ easy access to credits”. In the last chapter, I study the effect of privatization of state-owned enterprises in China, focusing on not only the effect of privatization on a firm's productivity, but also its effect on market output allocation. I use the method proposed by Olley and Pakes (1996) to find improvements in average market productivities for all industries in China. This growth in productivity resulted from enterprises becoming more productive, but not from more efficient output allocation within the market. Private firms are proven to be more productive than state-owned enterprises in all industries, but privatization itself improved firm efficiency only for some industries.

I study the impact of the used goods market on pricing and profits in the video game industry and the implications of resale restrictions. I develop a modeling framework that incorporates (a) heterogeneous consumers who are forward looking in their buying an selling behaviors, (b) a strategic game producer who prices its products considering both inter-temporal price discrimination and price competition with used goods, (c) rational expectations on future prices by both consumers and the firm, and (d) market equilibria for both new and used-goods markets. Without observing sales data, I use equilibrium pricing solutions in my model and the varying rate of price decrease after a game's release to identify the sales volume of a game in every period as a percentage of its total demand. I develop a computationally tractable utility specification to solve the computational challenge comes with modeling the supply side equilibrium. I construct the demand function for a game from heterogeneous consumers whose valuations distribute on an interval, and partially characterize the consumers' decisions and reduce the dimensionality of the state space. Applying the model to a unique dataset of game prices collected from the Internet, I estimate the game-specific demand for multiple games released in the U.S. market. The results show significant variation across games in terms of shapes of valuation distributions, expected play time, degrees of consumers' preference for new over used games, and price sensitivities. Policy simulations show that the effects of prohibiting resale largely depend on the shape of a game's demand distribution, because most of the profits are gained from higher-valuation consumers who purchase the game when the price is high. Prohibiting resale does not dampen their willingness to pay for the game because their high utility from playing it. Moreover, higher expected future prices in the absence of the used-game market further reduces their incentives to wait. I find the predicted profit increase is significant for most games when reselling is prohibited. However, games with demand consisting mostly of low valuation consumers benefit less from this structural change, because (a) early sales increase only slightly given a much smaller proportion of high valuation consumers and (b) losing the option to resell significantly decreases the willingness to pay for low valuation consumers, forcing the firm to slash its prices dramatically over time. I find empirical evidence that a firm can be better off with the used game market. This suggests that though eliminating the resale market is generally optimal for popular games, retaining it can be more profitable for some games. / text

The dynamic structural equation modeling (DSEM) framework incorporates hierarchical latent modeling (HLM), structural equation modeling (SEM), time series analysis (TSA), and time-varying effects modeling (TVEM) to model the dynamic relationship between latent and observed variables. To model the functional relationships between variables, a Gaussian process (GP), by definition of its covariance function(s), allows researchers to define Gaussian distributions over functions of input variables. Therefore, by incorporating GPs to model the presence of significant trend in either latent or observed variables, this dissertation explores the adequacy and performance of GPs in manipulated conditions of sample size using the flexible Bayesian analysis approach. The overall results of these Monte Carlo simulation studies showcase the ability of the multi-output GPs to properly explore the presence of trends. Also, in modeling intensive longitudinal data, GPs can be specified to properly account for trends, without generating significantly biased and imprecise estimates.

No description available.

Embora o Brasil tenha uma das maiores reservas de minério de ferro do mundo, estimada em 170 bilhões de toneladas, e também possua siderúrgicas próximas às minas de ferro, como a Usiminas, por exemplo, o Brasil exporta minério de ferro e importa produtos semi-acabados de aço de outros países, como o Japão e a China. O objetivo desta pesquisa é modelar a cadeia de valor global de ferro e aço e analisar sua dinâmica estrutural para verificar se esse modelo é capaz de determinar os fatores que impulsionam a geração de valor e que levaram alguns países a ter maior impacto sobre seu desenvolvimento econômico do que outros. A modelagem da cadeia de valor usa a técnica padrão de análise de entrada-saída de balanço de massa (método de modelagem estrutural ampliado de Leontief), detalhado na revisão da literatura. Para obter esses resultados, foi necessário coletar e consolidar diferentes bancos de dados utilizando ferramentas capazes de conciliar, tratar e analisar o grande volume de dados presentes nesses diferentes bancos de dados, em diferentes unidades e níveis de agregação. A análise da dinâmica estrutural revelou uma série de aspectos fundamentais para entender a guerra comercial entre os EUA e a China. A China passou de 10% da produção mundial de aço bruto para 50% em 20 anos, reduzindo a relevância de outros países no contexto da cadeia de ferro e aço. Além disso, a análise revelou que o Brasil exporta quase exclusivamente minério de ferro bruto. A Austrália, por exemplo, conseguiu agregar mais valor do que o Brasil. Isso revela, portanto, que a análise da dinâmica estrutural modelada por hipergrafos pode gerar informações semânticas relevantes sobre o contexto, atores e interesses envolvidos e pode servir como um mapa para orientar decisões e políticas voltadas para uma melhor inserção e atuação de um país em uma cadeia global de produção / Although Brazil has one of the largest reserves of iron ore in the world, estimated at 170 billion tons, and also steel mills near the iron mines, such as Usiminas, e.g., Brazil exports iron ore and imports (semi-)finished steel products from other countries, such as Japan and China. The objective of this research is to model the global value chain of iron and steel and to analyse its structural dynamics to verify if this model is able to determine the factors that drive the generation of value and that have led some countries to have greater impact on their economic development than others. Value chain modelling uses the standard mass balance input-output analysis technique (Leontief\'s extended structural modelling method), detailed in the literature review. To obtain these results, it was necessary to collect and consolidate different databases using tools able to reconcile, treat and analyse the large volume of data present in these different databases, in different units and levels of aggregation. The analysis of the structural dynamics revealed a series of fundamental aspects to understand the commercial war between the USA and China. China has shifted from 10% of the world\'s crude steel production to 50% in 20 years, reducing the relevance of other countries in the context of the iron and steel chain. In addition, the analysis revealed that Brazil exports almost exclusively crude iron ore. Australia, for example, was able to add more value than Brazil. This reveals, therefore, that the analysis of the structural dynamics modelled by hyper-graphs can generate relevant semantic information about the context, actors and interests involved and can serve as a map to guide decisions and policies aimed at a better insertion and performance of a country in a global value chain

This study investigated how social media use is related to body esteem (particularly weight satisfaction and muscularity importance) in adolescents. We conducted a 17-day Ecological Momentary Assessment study among 183 adolescents (12–17 years, 58% girls). Each adolescent reported on his/her social media use, weight satisfaction, and muscularity importance four times per day (68 assessments per participant; 6,863 completed in total). Using a person-specific, N=1 method of analysis (Dynamic Structural Equation Modeling), we found that at the between-person level, social media use is not associated with lower satisfaction with one’s weight and greater importance of looking masculine. At the within-person level, we found a significantly negative association for weight satisfaction and a significantly positive association for muscularity importance. For weight satisfaction, 76% of adolescents experienced no or very small effects as a result of SMU, 2% experienced positive effects, and 22% experienced negative effects. Regarding muscularity importance, 89% experienced no or very small effects of SMU on muscularity importance, 10% experienced positive effects, and 1% experienced negative effects. There is little evidence of gender differences in the effect of social media on body esteem in our sample.

Because little was known about how the masonry lighthouses constructed during the 19th century at exposed locations around the British Isles were responding to wave action, the dynamic response of the Eddystone lighthouse under wave impacts was investigated. Like other so called 'rock lighthouses', the Eddystone lighthouse was built on top of a steep reef at a site that is fully submerged at most states of the tide. Consequently, the structure is exposed to loading by unbroken, breaking and broken waves. When the breaking occurs, wave loading leads to complex phenomena that cannot be described theoretically due to the unknown mixture of air and water involved during the wave-structure interaction. In addition, breaking waves are generally distinguished from unbroken and broken wave due to the fact that they cause impulsive loads. As a consequence, the load effects on the structural response require a dynamic analysis. In this investigation the dynamic response of the Eddystone lighthouse is investigated both in the field and by means of a small-scale model mounted in a laboratory wave channel. In particular, field data obtained by the use of geophones, cameras and a wave buoy are presented together with wave loading information obtained during the laboratory tests under controlled conditions. More than 3000 structural events were recorded during the exceptional sequence of winter storms that hit the South-West of England in 2013/2014. The geophone signals, which provide the structural response in terms of velocity data, are differentiated and integrated in order to obtain accelerations and displacements respectively. Dynamic responses show different behaviours and higher structural frequencies, which are related to more impulsive loads, tend to exhibit a predominant sharp peak in velocity time histories. As a consequence, the structural responses have been classified into four types depending on differences of ratio peaks in the time histories and spectra. Field video images indicate that higher structural frequencies are usually associated with loads caused by plunging waves that break on or just in front of the structure. However, higher structural velocities and accelerations do not necessarily lead to the largest displacements of around a tenth of mm. Thus, while the impulsive nature of the structural response depends on the type of wave impact, the magnitude of the structural deflections is strongly affected by both elevation of the wave force on the structure and impact duration, as suggested by structural numerical simulations and laboratory tests respectively. The latter demonstrate how the limited water depth strongly affects the wave loading. In particular, only small plunging waves are able to break on or near the structure and larger waves that break further away can impose a greater overall impulse due to the longer duration of the load. As a consequence of the depth limited conditions, broken waves can generate significant deflections in the case of the Eddystone lighthouse. However, maximum accelerations of about 0.1g are related to larger plunging waves that are still able to hit the lighthouse with a plunging jet. When compared to the Iribarren number, the dimensionless irregular momentum flux proposed by Hughes is found to be a better indicator concerning the occurrence of the structural response types. This is explained by the fact that the Iribarren number does not to take into account the effects of the wide tidal range at the Eddystone reef, which has a strong influence on the location of the breaking point with respect to the lighthouse. Finally, maximum run up were not able to rise up to the top of the lighthouse model during the laboratory tests, despite this having been observed in the field. As a consequence, the particular configuration of the Eddystone reef and the wind could have a considerable bearing and exceptional values of the run up, greater than 40 m, cannot be excluded in the field.