Reaching for the stars: Australian firms and the global productivity frontier

Andrews, Dan; Hambur, Jonathan; Hansell, David; Wheeler, Angus Working Paper Reaching for the stars: Australian firms and the global productivity frontier Treasury Working Paper, No. 2022-01 Provided in Cooperation with: The Treasury, The Australian Government Suggested Citation: Andrews, Dan; Hambur, Jonathan; Hansell, David; Wheeler, Angus (2022) : Reaching for the stars: Australian firms and the global productivity frontier, Treasury Working Paper, No. 2022-01, ISBN 978-1-925832-41-9, The Australian Government, The Treasury, Canberra This Version is available at: https://hdl.handle.net/10419/280820 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. https://creativecommons.org/licenses/by/3.0/au/legalcode REACHING FOR THE STARS: AUSTRALIAN FIRMS AND THE GLOBAL PRODUCTIVITY FRONTIER Dan Andrews, Jonathan Hambur, David Hansell and Angus Wheeler1 Treasury Working Paper2 2022-01 Date created: January 2022 1 Jonathan Hambur is Macroeconomic Group. Dan Andrews, David Hansell and Angus Wheeler were all formerly in Macroeconomic Group, The Treasury, Langton Crescent, Parkes ACT 2600, Australia. Correspondence: [email protected]. We thank Peter Gal and Natia Mosiashvili for their comments and assistance with global frontier data. We also thank seminar participants at the RBA, OECD Global Forum on Productivity, as well as Rebecca Cassells, Nathan Deutscher and Mark Cully. 2 The views expressed in this paper are those of the authors and do not necessarily reflect those of The Australian Treasury or the Australian Government. © Commonwealth of Australia 2022 ISBN 978-1-925832-41-9 This publication is available for your use under a Creative Commons BY Attribution 3.0 Australia licence, with the exception of the Commonwealth Coat of Arms, the Treasury logo, photographs, images, signatures and where otherwise stated. The full licence terms are available from http://creativecommons.org/licenses/by/3.0/au/legalcode. Creative Commons BY Attribution 3.0 Australia licence requires you to attribute the work (but not in any way that suggests that the Treasury endorses you or your use of the work). Treasury material used ‘as supplied’ Provided you have not modified or transformed Treasury material in any way including, for example, by changing the Treasury text; calculating percentage changes; graphing or charting data; or deriving new statistics from published Treasury statistics – then Treasury prefers the following attribution: Source: The Australian Government the Treasury Derivative material If you have modified or transformed Treasury material, or derived new material from those of the Treasury in any way, then Treasury prefers the following attribution: Based on The Australian Government the Treasury data Use of the Coat of Arms The terms under which the Coat of Arms can be used are set out on the Department of the Prime Minister and Cabinet website (see www.pmc.gov.au/government/commonwealth-coat-arms). Other uses Enquiries regarding this licence and any other use of this document are welcome at: Manager Media Unit The Treasury Langton Crescent Parkes ACT 2600 Email: [email protected]ov.au Reaching for the stars: Australian firms and the global productivity frontier Dan Andrews, Jonathan Hambur, David Hansell and Angus Wheeler 2022#01 7 February 2022 ABSTRACT A feature of the recent global slowdown in productivity growth is that progress at the technological frontier has remained strong, while the gap between firms at the global frontier and other ‘laggards’ within an industry has grown. This growing gap reflects the fact that laggard firms now seem to be slower to adopt cutting-edge technologies and processes, and catch-up to the global frontier than they were previously. However, little is known about whether these patterns hold true for Australia. We exploit a novel dataset merging international microdata from OECD-Orbis with Australian microdata from BLADE. Consistent with overseas evidence, we find that the gap between global frontier firms and Australian firms has grown over time in the non-resource, non-financial market sector. Moreover, Australian firms catch up to the global frontier more slowly than previously, suggesting slower adoption of cutting-edge technologies and processes. The slowdown has been more notable in industries with declining measures of dynamism and competitive pressures, suggesting the slowdown may reflect weaker incentives and imperatives for firms to improve. This suggests that policies to address barriers to business dynamism and competitive pressures can improve Australia’s productivity performance, by increasing incentives for firms to adopt, innovate and improve. JEL Classification Numbers: C23, C55, D22, D30, E23, E24 Keywords: productivity, dispersion, firm-level, BLADE, frontier Jonathan Hambur Macroeconomic Group The Treasury Langton Crescent Parkes ACT 2600 1 1. INTRODUCTION Over the medium term, growth in Australia’s living standards will be driven by growth in productivity. However, like other advanced economies, productivity growth in Australia has slowed since the mid-2000s. Understanding the causes of this slowdown, and therefore which policies could address it, represents one of the most crucial roles for academics and policymakers. As such, over the past decade there has been a large amount of research and debate around the causes of the global productivity slowdown. One key question is whether the productivity slowdown reflects slower technological progress. For example, some argue that the productivity slowdown reflects a ‘return to normal’ following a period of transformative technological innovation (for example, electrification and information technology; Gordon 2012). Others argue that current innovations can be just as transformative (see Brynjolfsson and McAfee 2011; Mokyr 2013). Supporting this latter view, international evidence suggests progress for firms at the global frontier has remained strong. However, cutting-edge technologies and processes appear to be diffusing to other firms more slowly. In particular, Andrews et al (2019) find that the gap between global frontier firms and other ‘laggard’ firms within the same industries has grown, and laggard firms now seem to be slower to catch up to the global frontier. This suggests that the issue is slower diffusion of knowledge, not slower technological progress While these patterns have been documented internationally, little is known about whether they hold true for Australia. This paper exploits firm-level data from in Australia’s BLADE dataset and the OECD’s international Orbis dataset to provide new evidence on the labour productivity of Australian firms and how this compares to productivity for industry peers at the global frontier. We focus our analysis on the subset of non-financial non-resource market sector industries that are well-suited for benchmarking, consistent with the overseas literature. This covers manufacturing, goods distribution and retailing, and business and household services. However, we do exclude some sectors where Australian firms may be at the frontier, such as mining. First, we show that the productivity gap between the global frontier and Australian firms has grown over time. This pattern is consistent with evidence overseas, where firms across many countries have fallen further behind global leaders. In Australia, the divergence is particularly striking in the services sector, which is more protected from global competitive pressures. More importantly though, we find that Australian firms are catching up to the frontier more slowly than they did in the early 2000s. This suggests that Australian firms have been slower to adopt cutting-edge technology and processes, and to improve their productivity performances more generally. In turn, slower within-firm productivity growth has weighed on aggregate productivity growth. We then attempt to identify factors that could explain the slower catch-up, focusing on business dynamism and competition. We focus on dynamism because the entry and exit of firms intensifies competitive pressure on incumbents, forcing them to improve or exit. Entry also brings young firms into the market, who may be more likely to innovate and adopt new technologies or processes. Consistent with this expectation, firms in industries with higher entry, exit and turnover rates catch up to the global frontier more quickly, as do firms in industries with lower mark-ups (and therefore higher competitive pressures). Simple calculations suggest that the previously documented declines in business dynamism and competitive pressures therefore appear to account for half to three-quarters of the slowdown in the rate at which Australian firms catch up to the global frontier. 2 Overall, our findings suggest that part of the slowdown in aggregate productivity growth in Australia reflects slower diffusion of cutting edge-technologies and processes, and firm-level improvements, as in other advanced economies. They also suggest that, while global factors could be contributing to slower diffusion, policies that remove barriers to business dynamism and competitive pressure can encourage firms to catch up to the global frontier, thereby improving aggregate labour productivity performance. This motivates a greater focus on the potential obstacles to dynamism. Policies that facilitate more widespread adoption of emerging digital technologies can also play a role in improving productivity performance. The next section of this working paper discusses international evidence on the global productivity slowdown, models of firm catch-up, and market dynamism. Section 3 focuses on the datasets and the steps taken to clean the data, before presenting some preliminary findings. Section 4 sets up the econometric framework used to test for productivity convergence. Section 5 presents our key results before Section 6 concludes. 2. INTERNATIONAL LITERATURE Productivity growth has slowed significantly in advanced economies over the past two decades. While the exact timings and magnitudes differ across countries, productivity measures, and even studies, there is consistent evidence that productivity growth today is slower than it was previously. For example, comparing the periods 1996-2005 to 2006-2017, Goldin et al (2021) document declines in labour productivity growth of between 0.8 and 1.75 percentage points across five advanced economies. Given the centrality of productivity growth to long-run growth in GDP and living standards, a large literature has evolved trying to document the slowdown and understand its causes. 2.1. Techno-optimists and techno-pessimists A key strand of this literature relates to the pace and economic potential of current innovations. Techno-pessimists argue that the productivity slowdown reflects diminishing returns to new innovations. All the ‘low-hanging fruit’ innovations of the 19th and early 20th centuries, like electrification, were far more significant and transformative than anything seen since, and these innovations can only occur once (Gordon 2012). On the other hand, techno-optimists argue that current advancements, such as Artificial Intelligence, are just as transformative (Brynjolfsson and McAfee 2011; Mokyr 2013). How can this view latter view be squared with slowing productivity growth? One explanation put forward in the literature is that the process of technological diffusion has slowed. Technological diffusion is a key step in economic growth. Frontier firms innovate, and then over time other firms can adopt the new technologies, improving their performance and helping to further lift the productive capacity of the economy. This process of adoption means that firms further away from the frontier can improve productivity quickly and converge towards the frontier. Several empirical papers have found evidence supporting this notion of firm-level productivity convergence (for example, Andrews et al 2019; Griffith, Redding and Simpson 2009; Berlingieri et al 2020; Bartelsman et al 2008; Iacovone and Crespi 2010). But why would technological diffusion have slowed? One more benign explanation is that some recent innovations are ‘general purpose technologies’. These are technologies with broad use and application, which often take time to diffuse fully through the economy as they require substantial investments before they can be integrated fully into the economy. A past example is electricity; the process of electrification took decades even in more advanced economies (Bojanovic and Rosseau 2005). Another, less benign explanation is that structural factors that either prevent adoption – like financing frictions or anticompetitive use of intellectual property (Akcigit and Ates 2019) – or declining incentives to do so due 3 due to declining dynamism and less competitive pressure, have slowed the diffusion process (Andrews et al 2019). As can be gathered from the above discussion, a key aspect of the debate between techno-pessimists and techno-optimists is whether current technologies are inherently less transformative, or whether they are just diffusing more slowly through the economy. In the former case, the productivity frontier should be pushing out more slowly; in the latter case the frontier could still be growing quickly, but other firms will be slower to catch-up. Differentiating between these two explanations inherently requires firm-level data. 2.2. Growth of the frontier and productivity convergence A key recent piece of evidence in this debate was provided by Andrews et al (2019). They use cross-country firm-level OECD-Orbis dataset to examine productivity growth for firms at the frontier, and for laggard firms in the same industries. They find that progress for firms at the frontier has remained during the global productivity slowdown. This suggests that the techno-pessimist view is likely to be overstated. They also find that the gap between the global frontier firms and laggards in the same industry has grown, and that the rate at which firms converge to the productivity frontier has slowed. They argue that this suggests slower adoption and diffusion of cutting-edge technologies and processes throughout the global economy, and that this has weighed aggregate productivity growth. To understand the underlying driver of these patterns, Andrews et al (2019) explore the role of product market regulation. They find that in industries with less competition friendly regulatory regimes, the gap between the frontier and laggards tends to be larger. While the relationships between competition, and innovation and technology adoption is theoretically ambiguous (Aghion et al 2005), this provides some empirical evidence on the importance of competition in technology diffusion and adoption. Berlingieri et al (2020) similarly find evidence of slowing convergence across a number of advanced economies, though they focus on the national rather than the global frontier. Moreover, they find evidence that convergence tends to be slower where there are skill shortages/mismatches, such as in more digitally intensive sectors, and where financing constraints are more binding. Meanwhile, Akcigit and Ates (2019) examine the potential implications of slowing technology diffusion in a theoretical model with endogenous firm dynamics. Their results suggest that slower knowledge diffusion could explain the well-documented decline in business dynamism. Our paper adds to this existing literature by examining productivity convergence in Australia. We also contribute to the international literature by more formally examining the relationship between productivity convergence, and business dynamism and competition. 3. DATA AND GRAPHICAL RESULTS 3.1. Business Longitudinal Analysis Data Environment (BLADE) The analysis in this paper exploits firm-level data from the Business Longitudinal Analysis Data Environment (BLADE), compiled by the Australian Bureau of Statistics (ABS). BLADE captures administrative data from the Australian Taxation Office (ATO) for almost the entire population of 4 Australian firms, matched with ABS-produced survey microdata, such as the Business Characteristics Survey. We use data from firms’ Business Income Tax (BIT) forms, as well as their Pay-As-You-Go (PAYG) employment forms. BIT data are used to construct measures of gross value-added, defined as total compensation to labour plus gross operating surplus. Productivity is then measured as labour productivity, the ratio of gross value-added to full-time equivalent employees. We also use demographic information on firms contained in BLADE, such as age and industry. The focus of this analysis is on labour productivity, rather than multifactor productivity (MFP). This allows us to cover a larger sample of firms, both companies and unincorporated businesses, whereas we could only examine companies if we used MFP due to a lack of balance sheet and capital stock information for unincorporated business.3 Focusing on labour productivity may also be preferable given some technologies might be capital-embodied, and so abstracted from in using MFP. Still, future work could examine the results using MFP, which may be a cleaner measure of technological progress. As is common in the literature, we do not have access to firm-level input and output prices. As such we deflate value-added using industry deflators (1-digit ANZSIC). This is not ideal, as it means that our productivity measure will be affected by firm price differences, which could in turn reflect differences in market power, or product differentiation more generally. Nevertheless, numerous papers have shown that price- and quantity-based measures of productivity tend to move similarly, and failing to account for firm-level prices, while not ideal, does not have substantial impacts in firm-level studies of this type (Andrews et al 2019; Foster, Haltiwanger and Syverson 2008) 3.2. Data on the global productivity frontier Labour productivity data for the global productivity frontier come from the OECD-Orbis database. The database contains measures of labour productivity (similarly defined as gross value-added divided by full-time equivalent employees) at the firm-level and covers 24 OECD countries for the non-farm, non-financial business sector (Gal 2013; Andrews et al 2019). These data are sourced from annual balance sheet and income statements using a variety of underlying sources such as credit rating agencies, national banks and financial information providers.4 Labour productivity at the global frontier in each industry is taken to be the (unweighted) average labour productivity of the most productive 5 per cent of firms in that industry for a given year. Consistent with Andrews et al (2019), we identify the top 5 per of firms using a fixed number of firms across years. However, while the number of frontier firms is fixed over time, the set of frontier firms changes. This allows for churning at the frontier. The Australian frontier is defined in the same way using BLADE data. 3 See Hambur (2021) for a discussion. 4 For more information on the OECD-Orbis database, see Gal (2013) and the appendix of Andrews et al (2019). 5 3.3. Data cleaning and sample We take three main steps to clean the data and ensure comparability between the Australian and international datasets. First, we ensure comparability of nominal variables across countries and over time by adjusting for country-industry level differences in the purchasing power of currencies and applying industry-level deflators.5 Second, we convert Australian industry codes (ANZSIC 2006) to the NACE Rev 2 European classification system (which is equivalent to the international classification system ISIC Rev 4) using ABS concordances. Third, we clean the Australian data using the same methodology that Andrews et al (2019) use to clean the international data. This involves excluding outliers (defined as firms in the top and bottom 0.5 per cent of the labour productivity growth distribution), firms with less than three full-time equivalent employees, and any observations missing key information such as industry code, value-added and labour input. To be consistent with Andrews et al (2019), we confine our analysis to the market sector (that is exclude utilities, education, public administration and safety, arts and recreation, and health), exclude finance and insurance where productivity is notoriously hard to measure, and exclude highly volatile commodities sectors such as agriculture and mining. Quality data on the frontier are not available for many of these industries. We also remove construction due to difficulty measuring labour inputs given the use of contractors. This means retaining 2-digit NACE Rev 2 industry codes 10-33 and 45-82 (excluding 64-66), for which we have international comparator data. This covers just under half of gross value added in the Australian economy, and captures manufacturing, goods distribution and retailing, and business and household services. We confine our analysis to the 2002-2016 period, when we have both Australian and global frontier data. It is worth highlighting that Australian firms’ productivity performance may have been stronger in some of the excluded sectors. For example, many of Australia’s mining businesses are global leaders and may have had strong productivity growth over the period. After cleaning the data, the sample is an unbalanced panel of 1,372,576 observations, with an average of around 90,000-100,000 individual firms in each year. 3.4. Descriptive statistics As a first step, we plot unweighted-average productivity of firms in the global frontier, the Australian frontier, and Australian laggards, to understand of how Australian firms fair relative to the technological frontier. Figure 1 shows a growing gap between the productivity of the average global frontier firm and the average Australian firm. While productivity of those firms in the global frontier 60 per cent higher in 2016 compared to 2002, Australian frontier firms’ productivity is only 25 per cent higher, and other 5 To ensure monetary variables are comparable across countries, we adjust for differences in purchasing power of currencies. However, relative purchasing power across countries may differ by industry, for example as more developed countries often have higher services-goods price ratios. If this is the case, aggregate PPP conversions will not be sufficient to ensure comparability of labour productivity measures across countries. For this reason, we use industry-level PPP estimates compiled by Inklaar and Timmer (2014). To ensure comparability of monetary variables over time, we deflate the Australian data using ABS deflators at the 2-digit industry-level, and deflate the international data using 2-digit industry-level deflators from the OECD STAN database. 12 Table 4 – Convergence slowdown on separate samples Full Sample Mature Firms Young Firms (1) (2) (3) Distance from frontier 0.443*** 0.432*** 0.463*** (0.003) (0.004) (0.006) Distance from frontier x 2005-07 -0.009** -0.011** 0.000 (0.004) (0.005) (0.007) Distance from frontier x 2008-10 -0.005 -0.010** -0.018** (0.004) (0.005) (0.007) Distance from frontier x 2011-16 -0.039*** -0.047*** -0.001 (0.004) (0.005) (0.007) Industry -year fixed effects Yes Yes Yes Firm age and size controls Yes Yes Yes R -squared 0.209 0.197 0.243 Observations 1,372,576 996,829 375,747 Notes: Table 5 extends the baseline model by interacting the distance from frontier term with dummy variables for the 2005-07, 2008-10, and 2011 -14 periods. Column 1 runs this model on the full sample. Column 2 runs the model on a sub-sample of mature firms. Column 3 runs the model on a sub -sample of young firms. The coefficient on the distance from frontier term can now be interpreted as the convergence speed in the base period (2002 -04). Standard errors are clustered at the firm-level and are shown in brackets underneath each estimate. *significant at 10% level, ** significant at 5% level, *** significant at 1% level. This conclusion is also supported by re-running the earlier entry regression on a sub-sample of incumbent firms (Table 5). For these firms, convergence is slower when entry rates are lower, suggesting entry may affect convergence by intensifying competitive pressures on incumbents. Table 5 – Convergence and entry on separate samples Full Sample Incumbent Firms (1) (2) Distance from frontier 0.415*** 0.401*** (0.012) (0.012) Entry rate x distance from frontier 0.009*** 0.008*** (0.003) (0.003) Industry -year fixed effects Yes Yes Firm age and size controls Yes Yes R -squared 0.210 0.197 Observations 1,372,576 996,829 Notes: Table 6 extends the baseline model by interacting industry-level entry rates with the distance from frontier term. Column 1 runs this model for the full sample. Column 2 runs this model for the sample of incumbent firms. The distance from frontier and firm entry variables have been demeaned for ease of interpretation. The coefficient on the distance from frontier term can now be interpreted as t he convergence speed for firms in an industry with the average firm entry rate. The interaction term can be interpreted as the additional convergence speed associated with a 1 percentage point increase in the firm entry rate. Standard errors are clustered at the industry-level and are shown in brackets underneath each estimate. *signific ant at 10% level, ** significant at 5% level, *** significant at 1% level. To test this more formally, we incorporate a direct measure of competitive pressure, mark-ups as estimated in Hambur (2021). These measure the ratio of a firm’s sales price over their marginal cost of production and should capture the level of market power accruing to firms. They have increased over the past decade indicating decreasing competitive pressures in the Australian economy. 13 The firm-level mark-ups are aggregated to the industry level using either an unweighted average or a sales-weighted average. They are then interacted with the distance to the frontier, as was done with the measures of dynamism.8 Consistent with expectations, the rate of convergence declines as industries’ mark-ups increase, as evidenced by the negative and significant coefficient on the interaction between mark-ups and distance to the frontier (Table 6). This provides more direct evidence that declining competitive pressures can explain some of the slowing in convergence observed in Australia. Table 6 – Convergence and industry mark-ups Unweighted mark-ups Sales-weighted mark-ups (1) (2) Distance from frontier 0.349*** 0.390*** (0.039) (0.003) Industry level markup x distance from frontier -0.310*** -0.116*** (0.057) (0.019) Industry -year fixed effects Yes Yes Firm age and size controls Yes Yes R -squared 0.212 0.212 Observations 1,270,768 1,270,768 Notes: Table 5 extends the baseline model by interacting industry-level mark-ups with the distance from frontier term. Column 1 runs this model for unweighted averages of the firm mark -ups. Column 2 runs this model for sales-weighted averages of the firm mark-ups. The mark -up variables are demeaned for ease of interpretation. The coefficient on the distance from frontier term can now be interpreted as the convergence speed for firms in an industry with the average markup. The interaction term can be interpreted as the additional convergence speed asso ciated with a 1 percentage point increase in the firm entry rate. Standard errors are clustered at the industry-level and are shown in brackets underneath each estimate. *significant at 10% level, ** significant at 5% level, *** significant at 1% level. 5.3. Quantifying the effect of declining dynamism and productivity The previous section shows that in less dynamic or competitive industries firms appear to adopt new technologies and converge to the productivity frontier more slowly. Moreover, we know that measures of competitive pressures and dynamism have declined. So the natural question is, how much of the slowdown in productivity convergence can be explained by declining dynamism and competition? To consider this, we do a simple counterfactual. For each firm we calculate the implied convergence rate based on the earlier models and the observed mark-ups, or dynamism rate, as well as the rates implied by the model had mark-ups or dynamism remained at 2004 levels. We can then compare these two estimates: 𝐶𝐶𝐶𝐶𝑙𝑙𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝑙𝑙𝐶𝐶𝐶𝐶 𝐶𝐶𝑟𝑟𝑟𝑟𝐶𝐶𝑖𝑖,𝑖𝑖 ∆𝜇𝜇 =𝛽𝛽1+𝛽𝛽2∗ 𝜇𝜇𝑖𝑖,𝑖𝑖 𝐶𝐶𝐶𝐶𝑙𝑙𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝑙𝑙𝐶𝐶𝐶𝐶 𝐶𝐶𝑟𝑟𝑟𝑟𝐶𝐶𝑖𝑖,𝑖𝑖 𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶 𝜇𝜇=𝛽𝛽1+𝛽𝛽2∗ 𝜇𝜇𝑖𝑖,2004 The results of this exercise are shows in Figure 4. The decline in entry rates can account for around a 3 percentage points of the slowdown in the convergence rates, or around ¾ of the slowdown, while the increase in mark-ups can account for between 1¼-2 percentage point slowdown in the rate of 8 The main difference is that we also include a term for industry interacted with distance to the frontier. This accounts for the fact the level of mark-ups may not be well identified, though changes will be (see Hambur 2021 for a discussion). Not including this additional term does not change the results substantially. 14 convergence, or around ½-⅔ of the slowdown. While this is a simple exercise, it demonstrates that the decline in dynamism and competitive pressures have substantially lowered the rate of firm-level productivity convergence, and therefore productivity growth. Figure 4 – Drivers of the slowdown in convergence Notes: Figure shows total slowdown in convergence rate for 2011-2016 period based on Table 2 Column 2, as well as the portion of the slowdown explained by the entry rate, and mark-ups, using methodology discussed in section 5.3. Later component is shown for 2016, rather than for 2011-2016 average. 6. CONCLUSION This paper exploits a novel dataset merging international microdata from OECD-Orbis with Australian microdata from BLADE to analyse the performance of Australian firms relative to the global frontier. We show that the gap between the global frontier and Australian firms has grown over time, and Australian are catching-up more slowly. This suggests that Australian firms have become slower to adopt, innovate and improve their productivity performance, which can explain part of the slowdown in aggregate productivity growth since the mid-2000s. Similar, dynamics have been observed overseas Our results also show that slower catch-up and diffusion partly reflect declining business dynamism and competitive pressures. The motivates further research into the causes of the decline. From a policy perspective the results also suggest that policies that remove barriers to business dynamism and competitive pressure can encourage firms to catch up to the global frontier, thereby improving aggregate labour productivity performance. Policies that facilitate more widespread adoption of emerging digital technologies can also play a role in improving productivity performance. Understanding changes in productivity post-2016 and within the context of the COVID pandemic will be important research questions to pursue as data become available. Various data sources collected over this period have shown that Australian businesses have invested more in digitisation and innovative business practices, with the pandemic sparking an increase in technological adoption. This increased adoption of new technology will help Australian firms become more competitive in the global market and support future productivity growth. Worker mobility has also increased in recent times, leading to better matches between workers and employers. While much of this activity likely reflects delayed labour market movements due to the pandemic; continued labour market dynamism will also contribute to future productivity growth. 15 REFERENCES Aghion P, N Bloom, R Blundell, R Griffith and P Howitt 2005, ‘Competition and Innovation: An Inverted-U Relationship’, The Quarterly Journal of Economics, 120(2), pp 701-728. Akcigit U and ST Ates (2019), ‘What happened to US Business Dynamism?’, NNBER Working Paper No. 25756. Andrews D, C Cricuolo and PN Gal (2019), ‘The Best versus the Rest: Divergence across Firms during the Global Prodcutivit Slowdown’, CEP Discussion Paper No. 1645, August 2019. Andrews D and D Hansell (2021), ‘Productivity-enhancing Labour Reallocation in Australia’, The Economic Record, 97(317), pp 157-169. Bartelsman E, J Haskel, and R Martin (2008), ‘Distance to which Frontier?: Evidence on Productivity Convergence from International Firm-level Data’, Centre for Economic Policy Research Working Paper No. 7032. Berlingieri G, S Calligaris, C Criscuolo and R Verlhac (2020), ‘Laggard firms, technology and its structural and policy determinants’, OECD Science, Technology and Industry Policy Papers No. 86. Brynjolfsson E and McAfee A (2011), ‘Race Against The Machine: How the Digital Revolution is Accelerating Innovation, Driving Productivity, and Irreversibly Transforming Employment and the Economy’, Digital Frontier Press. Foster, L., Haltiwanger J and Syverson C (2008), ‘Reallocation, Firm Turnover, and Efficiency: Selection on Productivity or Profitability?’ American Economic Review, 98(1), pp 394-442. Gal PN (2013), ‘Measuring Total Factor Productivity at the Firm Level using OECD-ORBIS’, OECD Economics Department Working Papers No. 1049, OECD, Paris. Goldin I, P Koutroumpis, F Lafond and J Winklet (2021), ‘Why is Productivity slowing down?’, Oxford Martin School Working Paper Series on Economic and Technological Change No. 2021-6, May 2019. Gordon, RJ (2012), ‘Is US Economic Growth Over? Faltering Innovation Confronts Six Headwinds’, NBER Working Paper No. 18315, August 2012. Griffith R, S Redding and H Simpson (2009), ‘Technological Catch-Up and Geographic Proximity’, Journal of Regional Science, 49(4), pp 689-720. Iacovone L and GA Crespi (2010), ‘Catching up with the technological frontier: Micro-level evidence on growth and convergence’, Industrial and Corporate Change, Vol. 19, No. 6, pp 2073-2096. Inklaar R and Timmer MP (2014). ‘The Relative Price of Services’, Review of Income and Wealth, 60(4), pp 727-746. Hambur J (2021), ‘Product market power and its implications for the Australian economy’, Australian Treasury Working Paper No. 2021-03. Jovanovic B and PL Rousseau (2005), ‘General Purpose Technologies’, Handbook of Economic Growth, Chapter 18, Vol. 1(B), pp 1181-1224. 16 Ma S, J Murfin, R Pratt (2021), ‘Young firms, old capital’, Journal of Financial Economics, Online Access 24 September 2021. Mokyr J (2013), ‘Is technological progress a thing of the past?’, VoxEU column, 8 September 2013.