Research Interests


My research is at the intersection of labor and innovation aimed at understanding the way we work – how innovation impacts pay and the role that teamwork plays in the invention process.

I’m primarily interested in the rise of social, creative, and stem skills and its impact on wages and invention. In my previous work, I researched three topics: i. the impact of automation on inequality across Europe, ii. the role of teamwork (it’s impact on patenting throughout the life course), and iii. the importance of social & communication skills in knowledge diverse industries. My job market paper is on the latter topic.

Currently, I am an Assistant Professor of Economics at Pace University in New York, NY.

About Mary Kaltenberg

I am currently an Assistant Professor of Economics at Pace University. Previously, I was a post-doctoral fellow at Brandeis University with Adam Jaffe and Margie Lachman working on the inventor creative life cycle with patent data. I did my PhD at UNU-MERIT (Maastricht University), Maastricht, The Netherlands. My doctoral advisors were Bart Verspagen (UNU-MERIT), Neil Foster-McGregor (UNU-MERIT) and Cesar Hidalgo at the Collective Learning group at the MIT Media Lab (now at the University of Toulouse). In 2016, I was a visiting student and research assistant at the MIT Media Lab.

Previously, I worked at UNICEF on resource mobilization and research on accessibility to health care. I received my masters and bachelors degree in economics from The New School for Social Research in New York City (BA/MA program) .

Check out my Github for code and data from previous papers (currently updating), and code for tutorials for previously taught short courses.

I work on other personal projects/passions:

Funded Grants/Projects

  • SAFE (Safety Awareness for Empowerment): Development of an intelligent ChatBot for survivor resource assistance and NLP research on sexual assault (Supported by SWOL Limburg Fund and Diversity & Inclusiveness Grant, Maastricht University)
  • Check out the associated pre-print here. And some media mentions here and here (in Dutch)

Leadership and Community Service

Personal Projects

  • Cookbook, From Siberia to Texas: An Immigrant’s Collected Recipes
    Listen to my PechaKucha talk about it (currently writing a draft)
  • 3 time winner of the Red Sox dance off competition (2019)
  • Volunteer Cook for Community Cooks (2018-present)

Working Papers

Working Papers

The Knowledge Manager: Wage Premiums in Knowledge Diverse Industries (Job Market Paper)

Larger industries are known to pay more, but are these premiums simply a reflection of industry size, or are they an expression of increased knowledge diversity? Firms operate much like a team – they coordinate a variety of tasks and specializations to produce a good or service. In order to improve productivity, firms will seek to improve the efficiency of a task. Some ways that firms can do this is adopting a new technology, reduce coordination costs, and hire effective communicators. Firms that have to manage a wide variety of knowledge specializations have higher coordination costs, and these costs are especially high for firms that combine knowledge that are relatively dissimilar from one another. Firms that hire individuals who are effective communicators can reduce their burdening coordination costs, and are therefore willing hire individuals who have more social skills at a higher wage premium. I test this theory at the industry level, where I can empirically observe varying degrees of specialization. I develop a novel way to approximate knowledge specialization using an occupation-industry network. This network allows me to capture the variety of specializations in and industry, and the relative knowledge base distance between two occupations. At the industry level, workers who have social and communication skills will sort to industries that have a higher diversity of knowledge because those industries are willing to pay a knowledge diversity premium to reduce their high coordination costs. My results show that workers in industries with higher knowledge diversity receive a wage premium, especially for jobs with social, communication and interpersonal skills.

Reinventing the Elder Workforce: Creative Productivity Over the Life Course Through the Lens of Patent Data with Jaffe, A., and Lachman, M. (2019) In progress.

Previous research suggests creative ability peaks in the age decades of the 30s and early 40s, and declines thereafter, with some variation across fields. Building from the cognitive aging literature, we expect differences in the rate and quality of creative works by age. Cognitive processes show aging-related changes with increases in experience-based knowledge (pragmatics or crystallized abilities) and decreases in the ability to process novel information quickly and efficiently (mechanics or fluid abilities). We exploit a large database of U.S. inventors and the rich complexity of patent documents to explore both the rate and the nature of creative output over the life course. We extend this dataset by including information about age at patenting by combining public patent data with information on inventor ages scraped from directory websites on the web with approximately 1.2 million inventors patenting between 1976 and 2017. Our results suggest that cross-sectional and within-inventor patenting rates are similar, peaking at around the early 40s for both women and men. We find varying results for attributes of patents in relation to age some of which are consistent with cognitive aging theory. Backward citations and originality, which are connected to experience, were found to peak later in life. Forward citations and generality measures, which are more likely tied to fluid intelligence, peaked at earlier ages. These patent attributes also vary in relation to the age composition of the teams. Future goals are presented including consideration of the implications for work and retirement choices and policies.

Decomposing Inequality Across Europe: The Impact of Automation with N. Foster-McGregor

Existing research suggests that automation has the potential to impact employment and wage earnings. This paper focuses on the latter dimension and finds that the risk of automation has impacted wage earnings, and as a consequence has contributed to rising inequality in Europe. Using the structure of earnings survey (SES) we apply a RIF decomposition technique from Firpo, et. al., (2018) to uncover the drivers of the change in inequality between 2002 and 2014. The approach allows one to isolate the composition and the wage return effects of a variety of factors on the earnings distribution. We find that the characteristic that has the largest impact on inequality across all countries in our sample of European countries is the risk of automation. The impact of automation on inequality is found to be due largely to the composition effect, suggesting that workers are moving towards better paying low automation risk jobs, but the degree of wage dispersion between these jobs is higher than that for high automation risk jobs. These results point to evidence that the polarization effect of automation on worker earnings is occurring in many countries within Europe.

Local labor market and higher education mismatch: What is the role of public and private institutions?  with Ortiz, E.A., Jara-Figueroa, C., Bornacelly, I., and Hartmann, D. (2019) IDB Working Paper

Mismatches in the supply and demand of education and skills can hamper economic development. While most studies focus on skills mismatch at the national level, developing economies tend to have strong regional differences in their labor markets, and thus in the demand of skills. Yet few studies investigate to which extent enrollments in higher education match local labor markets demands. Moreover, there are few empirical studies comparing the effectiveness of private and public higher education institutions in coordinating the regional educational supply and demands. This is an important shortcoming in countries like Brazil where almost 90% of universities are private and strong levels of regional inequality and skills mismatches persist. In this article we analyze a large dataset of higher education enrollment rates in 22 different educational fields across 137 mesoregions of Brazil between 2010 and 2016. Our first stage regressions show that the local labor market conditions are a significant predictor of enrollment rates in different educational fields. Moreover, we identify relative levels of mismatches of each region. Less developed regions in particular show a high number of shortages in enrollments rates and thus supply of advanced skills for their labor markets. In the second stage regressions we analyze the factors explaining relative mismatches. The results show that a high level of urban population, low unemployment, and a large share of public universities are significantly associated with lower mismatches. In contrast, private universities do not significantly reduce, but rather tend to increase relative skills mismatches. In sum, our results imply that educational policies need to take the regional labor markets demand into account or they may perpetuate regional inequalities. Moreover, public universities in Brazil are more effective than private universities in promoting the enrollments in hard sciences and addressing the skills needs of less developed regions.

Mapping Stratification: the industry-occupation space reveals the network structure of inequality with Hartmann, D., Jara-Figueroa, C., & Gala, P. 2019. Working Paper

Social stratification is determined not only by income, education, race, and gender, but also by an individual’s job characteristics and their position in the industrial structure. Utilizing a dataset of 76.6 million Brazilian workers and methods from network science, we map the Brazilian Industry-Occupation Space (BIOS). The BIOS measures the extent to which 600 occupations co-appear in 585 industries, resulting in a complex network that shows how industrial-occupational communities provide important information on the network segmentation of society. Gender, race, education, and income are concentrated unevenly across the core-periphery structure of the BIOS. Moreover, we identify 28 industrial occupational communities from the BIOS network structure and report their contribution to total income inequality in Brazil. Finally, we quantify the relative poverty within these communities. In sum, the BIOS reveals how the coupling of industries and occupations contributes to mapping social stratification and the network structure of inequality.

Exporting Up: The Importance of Improving Technological Capabilities for Growth

What is the best way to measure technological capabilities? Over the past 15 years, technological capability indices have developed into two strains: aggregated capability indices and export based algorithms. We discuss the strength and weaknesses of using such measures and test at which point technological capabilities are important for low income nations to `catch-up’ with developed nations. We explore a variety of econometric estimation techniques including, random effect, fixed effect, Hausman-Taylor and GMM that compare three export based algorithms, economic complexity index, fitness and generalized fitness. Our results indicate that technological capabilities, measured with export based algorithms, contribute to economic growth for low income nations. However, we do not find conclusive evidence that these measures have an impact at all stages of the development process. We suggest that to understand how economic structures impact economic growth, future pathways of research should reevaluate how to measure complexity to include value added which is increasingly fragmented across global production chains, and to measure the complexity of service and knowledge based products which are becoming a pivotal part of economies across the world.

Other Working Papers

Albeaik, S., Kaltenberg, M., Alsaleh, M., and Hidalgo, C. A. (2017). 729 New Measures of Economic Complexity. arXiv preprint arXiv:1708.04107.

Albeaik, S., Kaltenberg, M., Alsaleh, M., and Hidalgo, C. A. (2017). Improving the Economic Complexity Index. arXiv preprint arXiv:1707.05826.

Jun, B., Kaltenberg, M. and Won-Sik, H. (2017). How Inequality Hurts Growth: Revisiting the Galor-Zeira model using the Korean Case. working paper.

Verspagen, B. & Kaltenberg, M., (2015). Catching-up in a globalised context: Technological change as a driver of growth, UNIDO Working Paper 20 | 2015.
Download it here or here

Technical Reports
Industrial Development Report 2016. The Role of Technology and Innovation in Inclusive and Sustainable Industrial Development. “Technological change, structural transformation and economic growth,” Vienna, UNIDO.

Hartmann, D., Jara-Figueroa, C. and Kaltenberg, M., 2017, The Brazilian Industry-Occupation Space: Structural Heterogeneity and the regional skills demand. IADB Technical Report.


Teaching Philosophy

I believe the best way to teach economics is through applied real world examples in an active environment. My pedagogical approach is to challenge students by asking questions and utilize classroom activities that reiterate the learning objective. This is especially important when teaching econometrics or statistics – the power of these tools is not clearly visible in theory alone, but also through application. Students also learn better when objectives are repeated in different ways – data collecting at the lecture, interactive websites, short videos at home, writing reports, and problems sets. Learning is best applied in a community of active and engaged students. My goal as a teacher is to foster this kind of environment.

Teaching Experience

Maastricht School of Governance, Master of Science in Public Policy

Introduction to Statistics (Fall 2014), TA

Introduction to Data Science (Fall 2015)
Feel free to request the syllabus and do files (Stata)

Introduction to Econometrics (Fall 2014 and 2015), TA

Intermediate Econometrics (Parallel Course to Intro to Econometrics) (Fall 2015), TA

Maastricht School of Governance, GPAC (PhD Program)

Intuition to Panel Data (Short Course 2017)
You can find the syllabus here
Feel free to request the do files (Stata)

Introduction to Stata (Short Course 2017)
You can check out the do files based on the course on my Github here

UNU-MERIT, (PhD Program)

Introduction to Python for Economists (Short Course 2017)
For jupytr notebook with the Python code, see my Github here

I am happy to provide teacher evaluations for all courses when available/applicable

Courses I never taught, but prepared a syllabus and hope it’s some use to the world:

Micro for a Digitized Economy (Graduate Level)
Applied Microeconomics – Digitized Economies

Applied Empirical Microeconomics of Economics (Graduate Level)
Applied Microeconomics



kaltenberg [at] merit [dot] unu [dot] edu
mkaltenberg [at] brandeis [dot] edu
mary.kaltenberg [at] maastrichtuniversity [dot] nl


University Address

Brandeis University
Brown 125
MS 062
Brandeis University
415 South Street
Waltham MA 02453-2728