RDP 2024-01: Do Monetary Policy and Economic Conditions Impact Innovation? Evidence from Australian Administrative Data 1. Introduction

Productivity growth is the key determinant of living standards over the medium term. Many factors can influence how quicky productivity grows, including technological change, competition, skilled labour, regulation, trade and tax policy (Aghion and Howitt 2008). But economists have traditionally assumed that productivity is unaffected by ‘cyclical’ factors such as current economic conditions and monetary policy, at least over the medium term.

More recently, though, there is a growing literature questioning this assumption. One stream of this literature argues that weaker economic conditions, particularly economic downturns, can lead to slower rates of innovation and technology adoption by businesses, which can in turn lead to persistent changes in productivity and economic output. These hysteresis effects are considered in Stadler (1990), Comin and Gertler (2006), Anzoategui et l (2019), Bianchi, Kung and Morales (2019) and Amador (2022) among others.

In a similar vein, there is evidence that contractionary monetary policy, which weakens economic conditions, can slow innovation and technology adoption, and therefore have persistent effects on productivity (Moran and Queralto 2018; Jordà, Singh and Taylor 2020; Ma 2023; Ma and Zimmerman 2023).^[1] While the effects of expansionary and contractionary monetary policy are likely to cancel out over a cycle, such a finding highlights the potential for medium-run economic scarring to occur if policy is constrained by the effective lower bound on interest rates and therefore cannot offset economic downturns.

This has important implications for macroeconomic policy, but there is little evidence on these issues outside the United States. And the effects could differ substantially in a small open economy like Australia that typically imports innovation, compared to a large economy like the United States that pushes the technological frontier. Moreover, much of the evidence has focused on narrow measures of innovative activity like research and development (R&D) spending and patenting. These measures are likely to miss a large amount of innovative activity in the form of adoption of existing technologies and processes. Empirically, adoption is an important determinant of aggregate productivity growth (OECD 2015; Argente et al 2020; Majeed et al 2021) and features in theoretical models of the effect of monetary policy on productivity (Moran and Queralto 2018).

We explore the effects of monetary policy shocks on several different measures of innovation in Australia, including broad measures that will capture adoption of innovation developed elsewhere. Further, we explore whether the effects differ by firm size and other firm characteristics, using firm-level information on innovation and adoption, which has received relatively little focus to date. As well as providing interesting insights into the heterogeneous effects of monetary policy, this helps us to better understand the mechanisms through which monetary policy, and economic conditions, could affect innovation. The channels we explore include: weaker demand, which lowers incentives to innovate; and tighter credit conditions, which make it harder to finance investment and innovation. We find that both are important. We also consider the effect of US monetary policy shocks, which could be important for a small open economy like Australia, both because of standard spillovers from US policy to domestic conditions (e.g. Georgiadis 2016; Kearns, Schrimpf and Xia 2023), and because Australia imports innovation from the United States and US policy affects US innovative activity.

We find that contractionary (expansionary) monetary policy shocks reduce (increase) aggregate R&D spending and that changes in R&D spending have medium-run effects on productivity. However, we do not find any effect of monetary policy shocks on the number of patents filed, consistent with Australia tending to be an importer of new-to-world innovations rather than a producer (Majeed and Breunig 2023).

Turning to the firm-level data, we find that overall monetary policy shocks appear to have relatively little effect on broader survey measures of innovation and adoption. But this result hides offsetting impacts of monetary policy by firm size. Following a contractionary policy shock the share of small and medium enterprises (SMEs) innovating declines, while the share of large firms innovating increases. These heterogeneous responses appear consistent with SMEs and large firms having differing exposures to the channels though which monetary policy affects innovation. Monetary policy tends to weigh on innovative activity by tightening credit constraints (the credit constraint channel), particularly for smaller firms. Monetary policy also weighs on innovation by lowering domestic demand (the demand channel), and exporting firms, who tend to be larger, appear less exposed to this channel. This latter result is flipped for US shocks, which, if anything, weigh more heavily on the innovative activity of exporters, who are more exposed to international conditions.

Overall, our results suggest that monetary policy and economic activity can have medium-run effects on innovation and therefore productivity, though they paint a more complex and heterogeneous picture compared to previous work. These effects are likely to cancel out over a cycle. However, the results reinforce the importance of using macro stabilisation policy to avoid sharp economic downturns, given the potential for medium-run economic scarring, and highlight the costs of having such policy constrained (Barlevy 2004; Benigno and Fornano 2018; Ikeda and Kurozumi 2019; Garga and Singh 2021). The potential for monetary policy to affect innovation and productivity in the medium run may also slightly alter the trade-offs between stabilisation of output and inflation when faced with a supply shock, particularly if inflation expectations can be kept anchored (Queralto 2022; Fornaro and Wolf 2023). That said, the results do not suggest that central banks should focus mainly on output growth at the expense of inflation stabilisation given the risks of expectations de-anchoring, which could require a much larger response and correspondingly sharper economic downturn.

This paper proceeds as follows. We first discuss the relevant literature on the determinants of innovation in Section 2 before giving an overview of our data and methodology in Section 3. We then look at the effects of monetary policy shocks on aggregate on innovative activity – patents, trademarks and R&D – and firm-level innovation in Section 4, before exploring the channels through which monetary policy may affect innovation in Section 5. We then consider the aggregate effect of changes in R&D spending on productivity in Section 6 before concluding.

Footnote

Baqaee, Farhi and Sangani (forthcoming) propose a different mechanism whereby contractionary shocks lead to a reallocation of resources towards lower productivity firms. [1]