Aniket Baksy

Between 1980 and 2000, growth in the skill premium and a decline in the relative price of capital led economists to conclude that capital-embodied technical change was driving up the relative demand for skilled labor. Given the continued steady decline in capital prices post 2000, these models predict a continual rise in the skill premium. However, post 2000, growth in the skill premium has slowed down. I argue that as the skill premium increased, firms adopted new technologies economizing on the use of skilled labor. I quantify this force using an equilibrium model with costly technology adoption. As capital prices fall, capital-skill complementarity initially drives up the skill premium. Firms respond by investing in new technologies which are less skilled-labor-intensive. The model successfully accounts for the slowing skill premium and the behavior of the labor share. Without technology adoption, the model predicts a skill premium in 2019 that is 5 percentage points higher and a labor share that is almost 12 percentage points higher. I provide microeconomic evidence for my mechanism by showing that accountants relatively more exposed to the adoption of accounting software saw slower wage growth.

We quantify the contribution of changes to the structure of the U.S. labor market to wage stagnation over the past 40 years. Using a rich structural model of wage and employment dynamics estimated on Current Population Survey data, we reach three main conclusions. First, upward job mobility has declined by half between the 1980s and 2010s. Second, this decline is not driven by weaker aggregate labor demand. Instead, we identify three key structural forces: increased mismatch between job openings and job seekers, rising employer concentration that limits job shopping opportunities, and reduced job search among the employed—potentially due to the growing use of non-compete agreements. Third, by curbing upward mobility, these structural shifts have lowered aggregate real wage growth by four percentage points since the 1980s, corresponding to approximately 40 percent of the decline of the aggregate labor share.

We examine patterns of economic policy uncertainty (EPU) around national elections in 23 countries. Uncertainty shows a clear tendency to rise in the months leading up to elections. Average EPU values are 13% higher in the month of and the month prior to an election than in other months of the same national election cycle, conditional on country effects, time effects, and country-specific time trends. In a closer examination of U.S. data, EPU rises by 28% in the month of presidential elections that are close and polarized, as compared to elections that are neither. This pattern suggests that the 2020 US Presidential Election could see a large rise in economic policy uncertainty. It also suggests larger spikes in uncertainty around future elections in other countries that have experienced rising polarization in recent years.

I argue that the successful incorporation of modern technologies into production processes requires large human capital investments. I argue that this can rationalize the divide between "digital superstars" who are enthusiastic early adopters of digital technologies and are reaping the productivity benefits, and small firms, who often make the case that new technologies are not useful in their businesses.

Although AI is increasingly applicable to business tasks, AI adoption remains low and concentrated in large firms, which increases inequality across firms and workers at those firms. We identify the key barriers to AI adoption as the high costs of AI customization to specific business needs of complementary data infrastructure needed to leverage AI. We propose two clusters of policies to lower AI adoption costs for small and medium enterprises (SME). First, we propose public support targeted at the creation and commercialization of flexible low/no-code AI platforms. Second, we propose creating public data repositories and a clearinghouse-like infrastructure to improve SME access to cutting edge pre-trained models and computational infrastructure. We also propose the creation of a medium-skilled data curator workforce to manage and reuse data and provide new opportunities for retrained workers.