CLSBE - Dissertações de Mestrado / Master Dissertations
URI permanente para esta coleção:
Navegar
Entradas recentes
- Stock market reaction to earnings announcementsPublication . Mondello, Alessandro; Revelo, José GarciaThis study examines how quickly earnings news is incorporated into prices in large U.S. equities and whether post-earnings-announcement drift (PEAD) remains economically meaningful in recent markets. Using quarterly earnings announcements for S&P 500 firms from 2022 to 2024, analyst expectations from I/B/E/S are matched to daily stock returns from CRSP and evaluated in a standard event-study framework. Earnings releases are classified according to whether reported earnings exceed, meet, or fall short of market expectations, allowing the analysis to separate “good news” from “bad news” events and to assess both the immediate announcement reaction and subsequent return continuation. The evidence shows that earnings announcements remain clearly value-relevant. Stock prices react in the expected direction of the earnings news, and the response is asymmetric: disappointments trigger larger and more pronounced adjustments than comparable positive surprises, while near-zero surprises generate little average movement. At the same time, PEAD is small and sensitive to the normal-return benchmark, and the post-announcement pattern is not monotonic in the sign of the surprise, appearing closer to mild mean reversion than to classic continuation. Extending the drift analysis back to 2010 indicates that PEAD has gradually weakened over time for large-cap firms, and the sector breakdown reveals meaningful heterogeneity in announcement effects across industries. Overall, the results are consistent with rapid price discovery around earnings releases in S&P 500 stocks, leaving only limited and fragile short-horizon return predictability in the post-announcement period.
- Cross-market arbitrage and market segmentation : evidence from high-frequency interest rate dataPublication . Clerici, Stefano; Cerqueiro, GeraldoThis thesis investigates the efficiency of financial markets by analyzing the segmentation between institutional Short-Term Interest Rate (STIR) futures listed on the CME and the emerging landscape of regulated prediction markets (Kalshi). While traditional finance theory posits that arbitrage forces should enforce the Law of One Price across functionally equivalent assets, we hypothesize that structural barriers and distinct participant compositions create persistent pricing discrepancies.Methodologically, we utilize a proprietary high-frequency dataset comprising tick-by-tick data for Fed Funds and SOFR futures, alongside limit order book data for binary event contracts. We first construct a rigorous "fair value" forward curve for the Effective Federal Funds Rate (EFFR), accounting for the term structure of calendar spreads and meeting probabilities. Subsequently, we develop a "hybrid hedging strategy" that utilizes a static replication approach to map the linear risk exposure of traditional futures onto the binary payoff structure of prediction markets.Our empirical results strongly validate the "Limits to Arbitrage" hypothesis. We document that while the institutional market operates with near-perfect efficiency, the prediction market exhibits systematic inefficiencies driven by retail behavioral biases (favorite-longshot bias) and one-sided liquidity constraints. By implementing a cross-market arbitrage strategy, we demonstrate the ability to generate risk-free alpha, confirming that market segmentation effectively insulates retail mispricing from institutional correction.
- Equity valuation of easyJet PLCPublication . Panzeri, Giorgia; Martins, José Carlos TudelaThis dissertation analyses the equity valuation of easyJet plc, a low-cost airline operating in the European market. The airline industry is characterised by strong competition, significant fixed costs and high exposure to economic cycles, which makes valuation particularly challenging. The objective of this study is to estimate the intrinsic value of easyJet and to compare it with the company’s market price. The analysis is based on a review of the company’s business model, operations and recent financial performance. Based on this analysis, financial projections are developed for the period from FY2026 to FY2030. These projections are used to value the company through a discounted cash flow model based on free cash flow to the firm. To support the results, the adjusted present value method and a relative valuation based on market multiples are also applied. The valuation also considers the company’s current investment programme, particularly fleet renewal, as well as relevant risk factors such as fuel prices, exchange rates and changes in demand. The discounted cash flow model estimates an equity value of £7.21 per share, which is higher than the market price observed at the valuation date. Overall, the results indicate that easyJet’s shares are undervalued.
- German energy-intensive firms and the Russia-Ukraine war : evidence of firm-level heterogeneity in crisis exposurePublication . Hirsch, Leonhard Peter; Cerqueiro, GeraldoThis study examines the stock market response of German energy-intensive firms to the Russia-Ukraine war. The objective is to determine if energy intensity predicts a different crisis exposure and to identify factors explaining within-sector heterogeneity. An event study methodology is applied to measure abnormal returns around the February 24, 2022 invasion, using 16 energy-intensive firms and 39 control firms listed on German exchanges. The results show no significant average treatment effect, however substantial firm-level variation exists within the treatment group, spanning a 77-percentage-point range. Steel and metals producers achieved returns between +10.9% and +21.0%, while chemical producers suffered losses between -10.9% and -23.3%. This deviation reflects whether firms use natural gas as substitutable energy input or non-substitutable chemical feedstock, a distinction that sector classifications fail to capture. The findings indicate that binary treatment-control classifications obscure important dissimilarity. The results are specific to German firms during the immediate outbreak period. Regardless, the evidence shows that firm-level measures of gas consumption intensity are necessary for accurately identifying crisis vulnerability.
- Interest rate effects on issuance and market activity of structured products in Germany & SwitzerlandPublication . Matos, Fábio Rodrigues; Schliephake, EvaThis dissertation investigates how interest rate changes influence the issuance and market activity of structured financial products in Europe, focusing on capital protection and yield enhancement products in Germany and Switzerland. Using quarterly data from 2012 to 2025, the study combines structured product turnover and issuance data with euro area and United States interest rates, equity market volatility indices, and exchange rates. The empirical analysis employs first-difference regression models with robust standard errors to capture short-term market responses.The results show that equity market volatility is the dominant driver of structured product activity across both countries and product types. Volatility has strong and statistically significant effects on turnover and, in some cases, on issuance, whereas interest rate effects are generally weak, unstable, and sensitive to model specifications. When interest rates matter, their effects differ by product family: capital protection issuance sometimes increases in higher-rate environments, while yield enhancement issuance tends to weaken as rates rise. A subsample analysis around the 2018 implementation of PRIIPs and MiFID II indicates that regulatory changes affected sensitivities but not the main conclusion. Overall, structured product activity is driven primarily by market uncertainty rather than by interest rate movements.
- Equity valuation of Costco Wholesale CorporationPublication . Kreisl, Paul; Martins, José TudelaThe objective of this thesis is to provide a fundamental equity valuation of Costco Wholesale Corporation using a combination of intrinsic and relative valuation methods. Costco operates in the mature and highly competitive retail industry, where scale, operational efficiency, and resilient business models are essential. Its membership-based warehouse format constitutes a key competitive advantage, generating strong customer loyalty, recurring revenues, and exceptional cash flow generation. The valuation combines a two-stage discounted cash flow model with trading multiples and precedent transactions. The resulting weighted fair value amounts to $489.86 per share. Relative to the market price of $938.82 on September 2, 2025, this implies a downside potential of roughly 48% and would correspond to a strong sell recommendation from a purely fundamental perspective. Although the intrinsic value lies significantly below the market price, this should not be interpreted as an indicator of an imminent correction. Instead, Costco’s valuation premium reflects the market’s willingness to capitalize the exceptional stability and resilience of its business model more strongly than current free cash flows alone would suggest. The company combines negative net debt with robust, largely cycle-resistant margins, fostering investor confidence in its ability to sustain strong operating performance over the long term. This helps explain why Costco trades at a valuation that meaningfully exceeds its intrinsic, cash-flow-based value estimate. A comparison with the Morningstar Equity Report shows that both analyses classify Costco as overvalued, although Morningstar arrives at a higher target price due to more optimistic assumptions regarding growth and profitability.
- Research report on Nvidia Corporation (NVDA)Publication . Moreira, Zacarias; Martins, José Carlos TudelaThis dissertation presents an in-depth valuation of Nvidia Corporation, analyzing how methodological choices, financial assumptions, and industry dynamics influence intrinsic value estimates. Grounded in academic valuation theory, the analysis applies a multi-method framework combining a discounted cash flow (DCF) model, a sum-of-the-parts (SOTP) valuation, and complementary relative valuation. Forecasts are based on a detailed assessment of Nvidia’s business model, competitive positioning, capital allocation strategy, and exposure to secular investment in AI infrastructure. The DCF model employs a six-year explicit forecast horizon, a bottom-up estimation of the weighted average cost of capital (WACC), and terminal values derived from both the Gordon Growth model and peer-based forward multiples. Scenario analysis highlights the sensitivity of valuation outcomes to long-term assumptions regarding data center capital expenditures, market-share evolution, and terminal multiples. The SOTP framework separately values the Compute & Networking segment and the remainder of the business, reflecting their distinct growth and margin characteristics. A comparison with a Zacks valuation report shows that divergences in target prices are primarily driven by methodological differences. While Zacks capitalizes near-term earnings using elevated forward P/E multiples, this dissertation discounts multi-year free cash flows at a double-digit WACC and applies more conservative terminal assumptions. Despite forecasting higher revenues and earnings, the resulting target price of $125.6 per share is materially below both the Zacks estimate and the current market price.The study concludes that Nvidia’s prevailing market valuation embeds optimistic assumptions regarding the sustainability of its growth, pricing power, and long-term returns on capital.
- Does the market perceive the Fed as independent?Publication . Ribeiro, Maria João Rodrigo Pereira; Stahl, JörgThis dissertation assesses the perception of the market regarding the Independence of the Federal Reserve via market participation. The Central Bank is responsible for the settlement of the Federal Funds Rate. Any shifts on this rate are decided by the Federal Open Market Committee, generating expectations on the population and market participants, essentially. The fundamental question arises: do investors genuinely trust in the institutional independence of the Federal Reserve? Such trust should exist: the credibility of this institution is a baseline for financial stability. However, it is also possible that investors do not, in practice, perceive the Federal Reserve as fully independent - examined in this study to infer investors’ underlying perceptions. This is studied through the behaviour of Fed funds futures contracts over specific historical events: events where it would be naturally expected that the Fed would intervene through shifts in the rates, due to purely sudden economic shocks or inflationary pressures; andevents where there is an external party synthetically trying to push the Fed to commit to a monetary policy decision. The study finds that the tendency of the market is a clear belief in Independence.
- Capital structure choices and crisis performance in U.S. REITs : the role of leverage, debt maturity, and property type differencesPublication . Rizzo, Alessio; Martins, AntónioThis dissertation examines how capital structure choices of U.S. Equity Real Estate Investment Trusts (REITs) before a crisis affect performance during two fundamentally different economic shocks – the Global Financial Crisis (GFC) and the COVID-19 pandemic. Since REITs have limited internal financing options due to statutory distribution requirements, they heavily depend on external financing sources, which makes the structure of leverage and debt maturities particularly important. Building on the framework of Pavlov, Steiner and Wachter (2018), this study analyzes performance effects through cross-sectional OLS and logistic regressions(“Does preparation matter?”), determinants of adjustment (“Who prepares?”), and if these relationships differ across property-type heterogeneity.The findings show that the effect of capital structure preparations is crisis-specific. While deleveraging before the GFC leads to significantly better crisis performance, neither leverage reductions nor adjustments of debt maturities had any impact on performance during the COVID-19 pandemic. Further, this study confirms that sectoral differences play a key role: Defensive REITs benefited significantly during the GFC with prior deleveraging, while during COVID, especially short-term debt maturities led to negative crisis performance for defensive REITs. Overall, this study shows that the effect of capital structure adjustments varies both across crisis types (credit market versus cash flow shock) and across different property type risks. Further, this study shows that crisis effects and property type heterogeneity jointly determine financial flexibility, which indicates that decisions over capital structure must be tailored to crisis and sector-specific conditions rather than applied universally.
- From patterns to profits : machine learning models for dynamic insider trading detectionPublication . Radivojevic, Nikola; Tran, DanThis thesis investigates whether machine learning models can extract profitable trading signalsfrom insider transactions disclosed in SEC Form 4 filings. Using 841,071 insider purchases andsales from 2010 to 2024 merged with CRSP market data, we train and compare three ensemblemodels: Random Forest, XGBoost, and CatBoost through time-series cross-validation with 34folds.Models achieve weighted precision of 78 to 79 percent, outperforming the 66 percent random baseline by 12 percentage points. However, accuracy (53 to 56 percent), recall (53 to56 percent), and F1 (61 to 64 percent) fall below the baseline, indicating conservative classification behavior. This precision-recall trade-off suggests models prioritize prediction accuracyover coverage, minimizing false positives at the cost of missed opportunities. Whether this conservative approach is acceptable depends on the application’s cost structure. Feature importanceanalysis reveals that portfolio turnover, stock price at filing, and market capitalization are themost informative predictors.Backtested trading strategies show that long-short portfolios generate statistically significantrisk-adjusted returns, with CatBoost achieving annualized alphas of 14.99 to 16.09 percentacross factor models and a Sharpe ratio of 1.69. Long-only strategies substantially outperformmarket benchmarks, while short-only strategies show negative overall returns but perform wellduring market downturns.However, high portfolio turnover (1,257 to 1,472 percent annually) generates substantialtransaction costs. Realistic execution costs eliminate the observed alpha, reducing economicviability particularly for retail investors. The findings confirm that insider trading signals contain exploitable information but emphasize the critical importance of execution quality and costmanagement for practical implementation.