From micro to macro: Using administrative and other microdata sources to answer macroeconomic questions

Abstract

This thesis presents four chapters using Australian microdata sources and techniques to explore important macroeconomic issues.

The first two chapters focus on productivity, the key driver of living standards over the long-term. The first of these explores the extent to which monetary policy, or demand shocks more generally, can influence innovative activity and productivity over the medium-term. While this has been explored for the US, it is yet to be considered for a small open technology importer, nor using broad measures of innovation that capture technology adoption, rather than just frontier innovation. Combining macroeconomic and firm-level measures of innovation with monetary policy shocks in a local projections framework, this work finds that contractionary policy can influence innovation. However, the effects are shorter-lived than previously found in the US and more heterogenous, with demand and financial constraint channels playing an important role in this heterogeneity. These findings reinforce the importance of macro-stabilisation policies to avoid medium-term productivity scarring

The second chapter examines the adoption of emerging digital technologies: Cloud Computing and AI/Machine Learning. Understanding barriers and enablers to profitable adoption of such technologies is crucial if we are to reap their benefits over coming years. Using private sector data, along with natural language processing, this Chapter builds a unique merged dataset of firm-level adoption, profits, hiring and management capabilities. Applying event study approaches, it finds that the average firm does not benefit form adoption in terms of profitability. But firms whose Board have relevant skills do. So this work highlights the important role of management skills in the profitable adoption of emerging technologies, which has implications for the focus of skills policies.

The third chapter focuses on inflation, and whether changes in the frequency of firm-level price changes can account for the pick up in inflation post-COVID. Using firm-level web-scraped prices microdata linked to firm tax data, the Chapter shows that firms began resetting prices more frequently in response to the large inflationary shocks post-COVID. This is counter to standard macroeconomic models which assume constant price-setting frequencies, and suggests the possibility of nonlinearities in inflation dynamics following large shocks. Using workforce policy DSGE model the Chapter shows that failing to account for these changes in price-setting leads to a significant underprediction of recent inflation outcomes. So central banks either need to invest in new frameworks, or add judgement during such periods.

The final chapter explores the labour market, and whether growing monopsony power caused slow wages growth in Australia over the 2010s. Using linked employer-employee administrative data it demonstrates that employer concentration is associated with lower wages. While labour markets did not become more concentrated, the effect of given levels of concentration appears to have increased. A key driver was the decline in firm entry, which meant there were fewer young growing firms trying to poach workers from incumbents. This work highlights the important links between firm dynamics and competition, and labour market dynamism , power and wages, and therefore between related policies. Show more