Working Papers

This thesis examines market-neutral, mean-reversion-based statistical arbitrage strategies in the Chinese equity market, using two factor decomposition methods: Principal Component Analysis (PCA) and sector-based Exchange-Traded Funds (ETFs). Residual returns are modeled as mean-reverting Ornstein–Uhlenbeck (OU) processes, generating contrarian signals. A 60-day rolling window ensures out-of-sample estimation. Realistic frictions are included via a 10-basis-point round-trip cost. Backtests from 2005 to 2024 compare four configurations: synthetic ETFs, fixed PCA, dynamic PCA, and trading-time volume adjustments. Both PCA- and ETF-based strategies deliver robust Sharpe ratios near 0.90–0.95. PCA portfolios perform better under high cross-sectional volatility, while ETF-based models remain stable during structural shifts. Incorporating trading volume enhances returns, especially for ETF models. Sensitivity analysis highlights the importance of threshold tuning and rolling-window lengths. These findings stress the critical role of factor construction and signal design in market-neutral strategies, suggesting further improvement via adaptive PCA and volume-weighted signals.

This paper investigates the profitability and robustness of pairs trading strategies based on non-parametric technical chart constructions—Renko and Kagi—across the U.S. and Chinese equity markets. Within a market-neutral, mean-reversion framework, the study examines strategy performance under varying market regimes, including the Global Financial Crisis (GFC) and the COVID-19 period. Using historical data from indices such as the S&P 500 and the CSI series, pairs are selected via statistical patterns in Renko and Kagi charts. Robustness checks consider varying trading horizons, the number of pairs, and transaction costs. Results show that both chart-based strategies generate significant excess returns and exhibit strong Sharpe ratios before costs. While trading frictions reduce profitability, Renko-based strategies remain resilient, especially during crises. The findings highlight that adaptive and non-parametric charting methods can effectively capture transient mispricings and provide viable alternatives for statistical arbitrage in turbulent markets.

This review examines the growing literature on pairs trading frameworks, which involve relative value arbitrage strategies between two or more securities. Existing research is categorized into five main categories: distance methods use nonparametric distance measures to identify pairs trading opportunities; cointegration methods rely on formal cointegration tests to reveal stationary time series of spreads; time series methods focus on finding optimal trading rules for mean-reverting spreads; stochastic control methods aim to determine the optimal portfolio holdings in pairs trading relative to other available securities; and the "Other Methods" category encompasses other relevant pairs trading frameworks, albeit with a more limited supporting literature. Through a comprehensive review of over 100 papers published between 2016 and 2023, the survey identifies the key strengths and weaknesses of each approach, providing insights relevant for future research and practical implementation.

This study investigates how work-from-home (WFH) —by mothers and their male partners—shapes maternal employment re-entry after childbirth. Drawing on Conservation of Resources and Boundary Management theories, the study distinguishes between WFH access and regular use. It hypothesizes that regular WFH use by mothers and their partners supports earlier and full-time maternal return to paid work, particularly among second-time mothers. The UK Household Longitudinal Study (2009–2019) is used to estimate discrete-time hazard models of return to paid work after first and second births, distinguishing between full-time and part-time re-entry. Among first-time mothers, both WFH access and regular use are associated with a greater likelihood of full-time re-entry, though not with overall return. Among second-time mothers, regular pre-birth WFH use significantly increases the likelihood of returning to paid work—regardless of hours—whereas access alone does not. No significant associations are found between male partners’ WFH and maternal employment outcomes.

This study explores how organizational factors influence managerial decision-making regarding the career advancement of employees working from home. Despite a large body of research on the new modes of working, a gap persists concerning the role of the organizational context in shaping these dynamics. In this article, we investigate whether managers’ promotion and pay decisions depend on the employee's use of remote work and whether these decisions are moderated by the presence of the ideal worker norms (i.e. high work devotion and centrality) and family-friendly policies (childcare-related and flexible work options) in their work environments. We use data from a choice experiment, which included over 1,000 managers from the United Kingdom. The experiment was run in the second half of 2022, and therefore, this study provides post-pandemic evidence and represents the “new normal” settings. The findings indicate that employees who work fully remotely are less likely to be considered for promotion and a salary increase than on-site workers. This pattern is observed particularly in firms with more demanding organizational cultures, namely those with stronger ideal worker norms and/or fewer family-friendly policies. Importantly, both male and female remote workers experience career penalties, albeit in distinct ways, as both ideal worker norms and family-friendly policies appear important for men, whereas for women, it is primarily the availability of supportive policies that influences outcomes. The findings underscore the significant impact of organizational culture on managerial decision-making, with implications for both theory and practice.

This study investigates the causal impact of prolonged school closures during the COVID-19 pandemic on non-cognitive predictors of mathematics achievement and the strength of their association with student performance. Drawing on data from TIMSS (2015, 2019, 2023) and PISA (2022), we apply difference-in-differences (DiD) models across two research designs: successive cross-sections of 4th-grade cohorts and a pseudo-panel following a cohort from primary to secondary school. Our findings indicate that, although school closures did not significantly affect the level of students’ self-beliefs, they did reduce the strength of the association between negative attitudes and achievement—particularly among girls and in OECD countries. The results highlight the nuanced effects of distance learning on mathematics outcomes, contributing to the literature on the role of affective-motivational factors in education.

Fertility rates in developed countries have declined to historically low levels, yet the reasons remain incompletely understood. While economic insecurity, housing affordability, and gender inequality have been central to earlier research, recent attention has shifted to climate change and reproductive rights. This study examines the relative importance of contextual constraints on childbearing intentions among young adults in Poland. Using a factorial survey (N = 1,337), we compare how employment stability, housing affordability, childcare availability, men’s domestic contributions, climate change, and abortion rights shape intentions     to have a first or second child. Results confirm the enduring salience of economic security for fertility intentions across gender and parenthood status. Reproductive autonomy is most influential for childless women and important for mothers and partnered men. Childcare and equitable domestic arrangements support women’s intentions but matter less for men. Climate concerns are less influential. Intentions of childless men appear less responsive overall.

The paper explores the use of Deep Reinforcement Learning (DRL) in stock market trading, focusing on two algorithms: Double Deep Q-Network (DDQN) and Proximal Policy Optimization (PPO) and compares them with Buy and Hold benchmark. It evaluates these algorithms across three currency pairs, the S&P 500 index and Bitcoin, on the daily data in the period of 2019-2023. The results demonstrate DRL's effectiveness in trading and its ability to manage risk by strategically avoiding trades in unfavorable conditions, providing a substantial edge over classical approaches, based on supervised learning in terms of risk-adjusted returns.

As labour-replacing technologies, such as industrial robots and AI-driven automats, alter the structure of labour demand, the task content of occupations has emerged as an increasingly important indicator of socioeconomic position. This study explores how exposure to job automation influences short-term marital intentions, using data from 1,345 respondents in the Czech Household Panel Survey (2015–2019) and occupational measures derived from the European Skills, Competences, Qualifications and Occupations (ESCO) Database. Our findings reveal a gendered pattern: men employed in highly routine-intensive occupations—indicative of greater vulnerability to automation—are more likely to report no plans to marry. Conversely, women in similar jobs are more likely to express positive marital intentions. These results highlight how technological change not only alters labour market outcomes but also shapes demographic behaviours through the lens of gender norms. The study contributes to broader debates on the interplay between socioeconomic disadvantage and family formation, suggesting that automation may be contributing to a growing group of ‘non-marriageable’ men. As technological change continues to transform the labour markets, especially with the rise of artificial intelligence, the number of individuals at risk of both economic and relational marginalization may further expand.
 

We examine the short-term fertility effects of Poland’s 2020 Constitutional Tribunal (CT) ruling, which declared abortions on the grounds of fetal anomaly unconstitutional. The decision effectively outlawed nearly all legal abortions, as over 97% had been conducted on this ground. Using vital statistics and interrupted time series analysis, we find a significant and immediate decline in births of around 6.6%. The fertility response was strongest among younger women and first-time mothers, suggesting heightened sensitivity to the increased risks of pregnancy. Contrary to expectations, highly educated women did not significantly adjust fertility, likely due to greater access to abortion services abroad. Our findings demonstrate that abortion bans may lower fertility when they substantially increase the perceived costs and risks of childbearing, particularly in societies with widespread contraceptive use. These results provide insights relevant to current debates in the United States, where overturning of Roe v. Wade may also reshape fertility patterns.

This pioneering study employs stated preference methods, specifically discrete choice experiments, to evaluate public preferences for the protection of diverse cultural heritage assets in Victoria, Australia. By analyzing responses to a series of hypothetical policy scenarios, we uncover the economic values the public assigns to various heritage attributes, including condition, accessibility, and protection measures. Our findings emphasize the importance of both use and non-use values in shaping willingness to pay for heritage conservation. These insights are critical for developing more effective, community-aligned heritage policies that reflect the public's valuation of cultural heritage. This research marks a significant advancement in the application of discrete choice experiments for general heritage valuation, offering a robust framework for future studies and policy development in cultural heritage preservation.

Understanding the value of cultural goods such as performing arts is essential for designing welfare-maximizing cultural policies. However, since these goods often generate public externalities, market data alone cannot capture their full value. While primary non-market valuation studies offer robust insights, they are resource-intensive. Benefit transfer (BT)—using value estimates from existing studies to assess unstudied contexts—offers a cost-effective alternative. This study examines the accuracy of BT in cultural economics, focusing on marginal values of theater performances. Using data from a discrete choice experiment (DCE) conducted in six Polish provinces, we perform inter-provincial transfers of marginal willingness-to-pay values for four types of theater performances. By comparing transferred values with primary estimates, we assess BT’s validity and reliability. To our knowledge, this is the first BT study in cultural economics using DCE-derived marginal values. Our findings suggest that BT accuracy in culture aligns with results from more established fields, supporting the use of DCE-based estimates in cultural policy evaluation.

This paper addresses whether women’s employment in the 21st century remains a barrier to family formation, as it was in the 1980s and 1990s, or—similar to men’s—it has become a prerequisite for childbearing. We address this question through a systematic quantitative review (meta-analysis) of empirical studies conducted in Europe, North America and Australia. We selected 94 studies published between 1990-2023 (N=572 effect sizes). Our analysis uncovers a fundamental shift in the relationship between women's employment and fertility. What was once a strongly negative association has become statistically insignificant in the 2000s and 2010s—and even turned positive in the Nordic countries and parts of Western Europe (France, Belgium, and the Netherlands). This shift is evident both among childless women and mothers and has occurred across all analyzed country clusters, except in the German/Southern European group, where the relationship has remained negative. These findings challenge longstanding assumptions about work-family trade-offs and suggest a reconfiguration of the economic and social conditions underpinning fertility decisions in contemporary high-income societies. The paper calls for a reconceptualization of the employment-fertility relationship and development of a new theoretical framework that better captures these evolving dynamics in contemporary high-income societies.

This paper examines the effects of asset purchase programs (APPs) that were implemented in a number of countries during the COVID-19 pandemic in concert with large fiscal stimulus packages. We identify APP shocks for 14 advanced and emerging market economies using high-frequency identification techniques. We next estimate panel local projections, finding that APPs tend to stimulate output, but decrease prices. By using a Kitagawa-Blinder-Oaxaca decomposition, we demonstrate that these responses significantly depend on the magnitude of the simultaneously applied fiscal stimulus. Remarkably, higher government purchases during that period crowded in private consumption and had a large effect on inflation. We show that these empirical findings, some of which are inconsistent with a standard New Keynesian framework, can be rationalized in a simple general equilibrium model with segmented asset markets and fiscal dominance.

We estimate the impact of railway construction on local populations in Russian Poland in the 19th century. The initial wave of railway expansion outpaced economic demand. From the late 1860s onward, locations connected to the railway network experienced significantly higher population growth. The economic effects of the connection to the railway network increased over time. State-funded military railway lines generated a smaller impact than private-owned lines. We also study the impact of the railway connection on social mobility proxied through a number of notable people born in a given city. However, we do not identify any robust impact.
 

This paper estimates the role of six shocks originating in the United States in driving the international business cycle. To this end, it employs impulse response functions and forecast error variance decomposition from panel local projections. We find that key macroeconomic shocks originating in the United States contribute significantly to business cycle synchronization between the US and other economies. These shocks also account for a substantial part of output fluctuations in these economies. Considering individual shock contributions, we document that technology and monetary policy innovations are of the highest relevance.

This paper investigates the domestic and international transmission of U.S. fiscal news shocks emphasizing the importance of the sentiment channel for the global economy. We identify these shocks using federal government spending forecasts from the Survey of Professional Forecasters. Employing the local projection method, we find  that anticipated increases in U.S. government spending are expansionary domestically, leading to improved sentiment and enhanced financial conditions. On the other hand, the U.S. dollar appreciates, and the U.S. trade balance deteriorates when future fiscal expansion is expected. In the international context, we apply panel local projection models across a broad set of countries and show that positive sentiment and improved financial conditions driven by U.S. fiscal news spill over, stimulating demand and output growth in other economies. However, we find no significant effect of currency depreciation on net exports in a broad sample as rising domestic demand tends to boost imports. In turn, in a subsample of countries with high trade exposure to the U.S., the trade channel becomes significant, while financial channel diminishes in importance. At the same time, sentiment channel appears to play a significant role in all subsamples. Finally, we find that positive fiscal news shocks have strong stimulating effects (both domestic and international) during US recessions but not in expansions.

The theory of disruptive innovation has gained significant interest from the academic and business communities, investigating the reasons for inadequate or delayed responses of incumbents to innovation which challenges their established markets. Inadequate responses were explained by the concept of innovator’s dilemma, which postulates how incumbents tend to satisfy their consumers on the established markets, while overlooking opportunities with other consumers. However, the optimal incumbents’ response to disruptive innovation in the normative sense has not been researched. We investigate the choice of the payoff-maximizing strategy in response to the observed market disruption as a choice between different management approaches. Our study shows the importance of the speed of innovation adoption for the choice of the optimal response to innovation. We also show that the response can be delayed in some cases, reflecting the exploitation-exploration dilemma. This new insight complements the reasons of delayed response to disruptive innovation threats.

We provide the first estimate of the Polish national, regional and sectoral GDP in the interwar period. We find that the Polish economy's performance in the interwar period was much better than it was assumed before. In the years 1924 – 1938, the real GDP per capita increased by almost 40% or by 2.3% annually. As economic growth was stronger in the poorer regions significant regional convergence was achieved. Our results challenge the dominant narrative about the weak performance of the Polish economy in the interwar years.

We measure the value added in agriculture in Poland during the interwar period. Our calculation is based on the bottom-up methodology. We provide estimates on the national and regional levels. Cultivated area, yields and yields per hectare increased during the investigated period. Significant regional convergence, both in the case of prices and value added occurred. In the years 1924 -38 value added increased by 5.35% annually, resulting 4.01% per capita growth rate. However, the yields per hectare grew less than in a majority of other European economies. While less developed eastern regions caught up with more economically advanced western Poland, the leading west lost compared to European peers. Therefore, our assessment of the development of agriculture in Poland in that period remains mixed.