1、1SEPTEMBER 2024TECHNOLOGY SPENDING:DRIVING GROWTH AND RESILIENCE THROUGH ECONOMIC DOWNTURNSin association with CONTENTSEXECUTIVE SUMMARY.4 Surviving and thriving in the next downturn:The right technology investments pay off.4Key findings.5SECTION 1:Introduction.7Box 1:A unique technology spending da

2、taset.9SECTION 2:Evolution of tech spend.10Tech spend has outpaced GDP and other investments.11Innovative disruption has favored certain types of tech spend more than others.12SECTION 3:Tech spend as a driver of business outcomes.14More than just a strong correlation between tech spend and revenue g

3、rowth in normal times and during downturns.15Tech spend has been important for resilience and recovery.15Box 2:Review of the dynamic panel model used in this study.16Data deep-dive reveals four key takeaways for decision makers.18SECTION 4:Conclusion.22SECTION 5:Areas for further investigation.24App

4、endix 1:Tech spend during recessions.28Appendix 2:Industry-specific analysis.31Appendix 3:About our econometric analysis.37ABOUT OXFORD ECONOMICS.38Table of contents 34 EXECUTIVE SUMMARYSURVIVING AND THRIVING IN THE NEXT DOWNTURN:THE RIGHT TECHNOLOGY INVESTMENTS PAY OFFHow organizations invest in te

5、chnology during economic downturns can be a strong indicator of their performance during hard times and the recoveries that follow.Conventional wisdom holds that cost cutting is an essential way to survive a recession,and while that may hold true in some areas,new research from EY and Oxford Economi

6、cs shows that strategic investments in technology during a downturn may be the leading strategy for surviving a recessionand thriving in a recovery.That does not mean organizations should blindly throw money at technologyas many companies learned during the Covid-19 era,it is possible to overspend e

7、ven when real opportunities exist.Much of the frenzied buying during the early 2020s resulted in a glut of technology assets that brought costly complexity to computing environments and created a tech hangover that has had a negative effect on at least short-term spending.Nor does spending through a

8、 recession mean that there is a single correct way to approach technology investments when the economy goes south.One key to a successful strategy is understanding the nature of the situationnot all downturns are created equal.Over the last three decades,each economic crisis has had unique causes an

9、d effects,and disparate impacts,on particular industries.There is no one-size-fits-all model for technology spending.When EY and Oxford Economics teamed up for this research,we found little empirical evidence of technology investment strategies during downturns that could enable robust guidance to c

10、orporate decision makers.Much of the existing analysis focuses instead on individual business cases or specific products.Against this background,EY and Oxford Economics compiled a significant new dataset of technology spending from publicly available industry data to extract key lessons for investme

11、nt strategies across a broad spectrum of industries and product categories in advanced economies,using econometric analysis combined with insights from academic literature and business leaders.One key to a successful strategy is understanding the nature of the situation not all downturns are created

12、 equal.5 Chart 1:Two-year avg.annual growth gap in industries with 1 ppt lower tech opex spend as a share of revenuesKEY FINDINGSOur research shows that,in broad terms,industries that spend wisely on technology through economic downturns emerge stronger and more competitive while avoiding the worst

13、of the technology hype cycle.Among the core messages:Recent analysis of macro trends confirms the increasing significance of technology spending in all economic sectors.This is evidenced by businesses committing an increasing share of revenues to tech spend over time.The surge is more pronounced in

14、services industries,where increased tech spend has become a crucial driver of economic activity.Compared to producing industries,service sectors are witnessing a more substantial productivity boost from technological advancements.Importantly,newer studies tend to find a higher importance of software

15、 and cloud-related products(preferred by services industries),while older studies tend to find a bigger impact of hardware and communications technologies.Tech spending can support revenue resilience during a downturn.On average,revenues shrank by 3.5 percentage points less in industries entering th

16、e 2008 global financial crisis(GFC)in the top quartile in terms of relative tech spend.And during Covid-19,industries in the middle of the pack averaged revenue losses 8 ppts larger than those at the top of tech spending.Notably,regression results suggest that shifts in the composition of tech spend

17、 play an important role in downturn resilience.In the near term,reducing the level of operational expenditures(opex)relative to revenues has been considerably harmful to revenue growth during downturns,suggesting that this could be a sub-optimal strategy for protecting profitability in times of cris

18、is.By contrast,changes in tech capital expenditures(capex)had limited impacts on growth,flagging that many industries may have overspent during the GFC and Covid-19 crisis.The bars in the chart illustrate the estimated gaps in revenue growth between industries with different levels of tech opex acro

19、ss the entire sample and during downturns.The gray bars show the gap over the entire sample,and the red bars show how much bigger the gap becomes during a downturn.These regression coefficients essentially quantify average outcome differences across industries and countries.Technology expenditures a

20、re normalized and analyzed as a share of revenues.To account for sector-specific performance,these figures were estimated using Oxford Economics dynamic panel model econometric framework discussed in Box 2 and in the technical appendix.5Impact on revenue growth(percentage points,2yr avg.)Last 3 down

21、turnsCOVIDBase impactAdditional recession specific impact-0.4-0.3-0.2-0.10Source:Oxford Economics estimates6 Tech spending fuels recovery.Investment in tech before and during a downturn can do more than just keep businesses afloat;it can carry them through and hasten recovery.Excluding heavy manufac

22、turers and public services,average post-downturn revenue growth within a parent sector was between 1.2 and 3.8 ppts higher in industries that maintained a relatively high level of tech spend during the GFC recession.1 On average,industries that go into downturns with larger shares of tech spend also

23、 appear to be better placed to seize growth opportunities after a recession.Our regression results also point to a consistent level of tech opex(pre-and during a downturn)as key factors for post-downturn competitiveness.By contrast,capital expenditures(capex)in tech were not found to have an immedia

24、te impact on resilience,but in some industries there were indications they could be important drivers of growth 24-36 months ahead,especially after recessions.Know your downturn,and target technology based on specific needs.Different lessons apply in different types of downturns,whether a tech stock

25、 market crash,a global financial crisis,or pandemic-driven shutdown of the world economy.Although spending through a downturn may become the new normal,that does not translate into throwing money at technology.Understanding the macroeconomic context,including political and socio-cultural trends as w

26、ell as the source of the shock,is critical to making the right investment decisions.A financial crisis that reduces credit availability implies a different set of technological investments than a pandemic that causes people to stay home and socially distance.Further,those same shocks imply a differe

27、nt set of investments for companies operating in different sectors.1 For comparison purposes,the granular industries covered in this study are grouped into six parent sectors:private services,public services,retail and wholesale trade,transportation and storage,heavy manufacturing,and light manufact

28、uring.A financial crisis that reduces credit availability implies a different set of technological investments than a pandemic that causes people to stay home and socially distance.7SECTION 1:INTRODUCTION78Section 1:Introduction2 Stiroh,K(2005).Reassessing the Impact of IT in the Production Function

29、:A Meta-Analysis and Sensitivity Tests.Annales dconomie et de Statistique No.79/80,pp.529-561.Available at https:/doi.org/10.2307/207775873 Spiezia,V(2012).ICT investments and productivity:Measuring the contribution of ICTS to growth.OECD Journal of Economic Studies,Vol.2012/1.Available at http:/dx.

30、doi.org/10.1787/eco_studies-2012-5k8xdhj4tv0t 4 OECD(2015).The Impact of R&D Investment on Economic Performance:A Review of the Econometric Evidence.OECD Directorate for Science,Technology and Innovation.Available at https:/one.oecd.org/document/DSTI/EAS/STP/NESTI(2015)8/en/pdfINTRODUCTIONAlthough t

31、here is a wealth of research demonstrating the positive and significant effects of technology investments on productivity and revenue growth,2,3,4 a relatively common strategy encourages companies to reduce their spending on technology to protect profitability during economic downturns.As shown in C

32、hart 2,this is the main reason business spending on tech has exhibited a cyclical pattern over the past 25 years,with significant fluctuations in growth.There is also evidence of scarring,or long-term damage,to technology spending following economic downturns.The analytical work presented in this pa

33、per reveals that this approach has been short sighted and has ultimately harmed long-term business success.Industries across different countries have exhibited varying responses to economic downturns,and our empirical study demonstrates the significant impact of these differences on revenue growth.I

34、n particular,the granularity of our unique technology spending dataset(described in Box 1)allowed us to uncover unique insights about tech spend and its composition on resilience,post-downturn competitiveness,and innovation.For example,our dataset reveals that,on average,companies in goods-producing

35、 industries have cut their technology spending three times more than those in service industries during challenging economic times.After controlling for sector-specific performance constraints,our econometric results show that this has typically contributed to lower revenue growth during the last th

36、ree global recessions and in some instances has also constrained post-recovery performance.Chart 2:Global tech spend growth-12-60612202020182016201420122010200820062004200220001998%growth,constant pricesSource:Oxford Economics estimates9Section 1:IntroductionHow technology investments enable growth

37、and competitiveness Two key factors help explain the strong relationship between technology spending and growth.Technology is a driver of innovation and competitiveness in the modern global economy:Companies that invest in technology are more capable of adapting to changing market conditions,enhanci

38、ng their products and services,securing market access,and staying ahead of their competition.Reducing technology spending during a downturn can impede a companys ability to innovate and compete,ultimately resulting in decreased profitability.Technology helps maintain and improve operational efficien

39、cy:By investing in new technologies,companies can optimize their processes,lower costs,and increase productivity.Reducing technology spending during a downturn can consequently lead to decreased operational efficiency,resulting in higher costs and lower profitability over time.Of course,it is possib

40、le to overspend on technology.Companies must strike a balance between investing enough to remain competitive and innovative,while avoiding certain types of technology that may not provide a sufficient return on investment.It is crucial for business leaders to carefully assess the composition of thei

41、r technology spending against relevant industry benchmarks.Overall,lessons from our study are intended to serve as a guideline for executives to find the optimal course of action given company characteristics and where they are in the business cycle.9BOX 1:A UNIQUE TECHNOLOGY SPENDING DATASET This s

42、tudy is based on an original dataset of technology spending put together by Oxford Economics in 2023 to provide valuable insights into the impact of different types of technology investments on company revenues and resilience during economic slowdowns and downturns.A globally consistent,industry-lev

43、el dataset for a sufficiently large number of countries is essential for hypothesis testing to deduct robust claims regarding the association of technology spending with performance.The dataset includes information from 11 countries and up to 35 industries per country,spanning from 1995 to 2021.Five

44、 core enterprise technology items are considered along with two additional drivers of enterprise value,which we consider part of the broader “tech spend”umbrella.These are the following:Devices:Electronics,computers,communication equipment Enterprise software:“Pre-packaged”software IT services:Custo

45、m and“own-account”software and related services Internet and cloud services:Data processing,internet portals,search engines Communication services:Wired,wireless,satellite Research and Development(R&D):Innovative activities to generate new products Brand management:Advertising,media,and public relat

46、ion systems,artistic originals For each,two additional breakouts are included:Capex(capital expenditure):Long-term expenditure related to plans for future growth and efficiency,that are purchased for a life span longer than one year(e.g.,purchases of IT equipment,enterprise software required to run

47、production facilities)Opex(operating expenses):Day-to-day expenditure related to each industrys production requirements that are purchased for use in making the industrys product(e.g.,annual software subscription,semiconductors installed in cars)The data is compiled from publicly available sources,p

48、rimarily from organizations such as the OECD,the US Bureau of Economic Analysis,and the Australian Bureau of Statistics.SECTION 2:EVOLUTION OF TECH SPEND1011Section 2:Evolution of tech spendA broad review of existing literature and a brief look at our technology spending dataset shows how tech spend

49、 has evolved over the past 25 years,with particular focus on economic downturns.KEY FINDINGS AND TAKEAWAYS Data reflects that tech spend has become ever more important across all sectors of the economy,outpacing revenue growth over timeespecially for capex.The rise has been particularly notable for

50、service industries,which may see a larger productivity benefit from tech compared to producing industries.Tech spend has responded differently during different downturns.After a considerable tech hangover in the early 2000s,the belief that tech boosts productivity and profitability has driven strate

51、gy even during downturns,with a spending peak during the Covid-19 pandemic.TECH SPEND HAS OUTPACED GDP AND OTHER INVESTMENTSThe digital revolution has supported a considerable rise of technology-related expenditure over the past 25 years,outpacing both total investment and GDP(Chart 3)hinting at the

52、 rising importance businesses assign to tech spending as a catalyst for growth.The rise has been particularly notable for service industries,which may see a larger productivity benefit Chart 3:World GDP and investment1997200020032006200920122015201820210100200300GDPTotal InvestmentTech spendIndex(19

53、97=100),inflation adjustedSource:Oxford Economics estimates12Section 2:Evolution of tech spend20002005201020152020Heavy manufacturingLight manufacturingPrivate servicesPublic servicesTradesTransportation and storage2000200520102015202020002005201020152020200020052010201520202000200520102015202020002

54、005201020152020012340246048120240240123%from tech compared to producing industries(Chart 4)the relative share of tech spend as a percentage of revenues went from 6%before 2000 to over 9%after the pandemic.Chart 4:Tech spend(percent of revenues)INNOVATIVE DISRUPTION HAS FAVORED CERTAIN TYPES OF TECH

55、SPEND MORE THAN OTHERSThe share of revenues that industries spend on enterprise software has grown six-fold since 1997,reaching a global average of 0.6%in 2021.This reflects the increasing reliance of businesses on software solutions to enhance their productivity,innovation,and competitiveness.Howev

56、er,software is not the only category that has seen a significant increase in spending.Internet and cloud services,which include data storage,processing,and analytics,have also grown rapidly,especially in the last few years.In fact,the share of revenues spent on internet and cloud services has almost

57、 tripled since the GFC,and is now close to matching software spending(see Chart 5).By contrast,communication services have experienced a steady decline.The shift toward mobile devices and the increased use of cloud-based services(e.g.,VoIP)have likely hurt communication services;however,as we explor

58、e in this study,communication services are still an important driver of revenue growth.What explains this shift in technology spending?One of the key factors is the rising importance of intangible assets,such as data and computer servers,in business operations.These assets enable businesses to lever

59、age the power of cloud computing,artificial intelligence,and other emerging technologies to create value and gain competitive advantage.Our econometric analysis,which is available in the appendix of our report,provides empirical evidence for this relationship.However,the data we use in our study doe

60、s not fully capture the complexity and Source:Oxford Economics estimates13Section 2:Evolution of tech spenddynamism of the technology landscape because standard industry classifications are based on historical definitions used by statistical agencies and may not reflect the current and future realit

61、ies of technology usage.For example,cloud services are a relatively new phenomenon that span across multiple categories,such as software,IT services,and internet services.As such,they may not be adequately represented by any single category in our data.Therefore,we recommend that more granular and t

62、imely data sources be developed and used to better understand the strategic implications of shifting from one type of technology to another.Chart 5:Global tech spend evolution by type(percent of revenues)How recessions shaped tech spendAs might be expected,the nature of the downturn affected tech sp

63、endboth in the volume of spend and the types of technology targeted.During the dotcom crisis,IT was viewed with skepticism by executives who cut their technology budgets to survive the downturn.Tech spending entered a period of austerity.However,by 2008,executives began to realize that technology wa

64、s no longer a“nice to have”but a necessity.The types of tech spend cuts have also varied significantly after each of the past two recessions.For example,global devices spend took a hit during the GFC but were boosted during the Covid pandemic,which also saw a rise in investments in digital infrastru

65、cture.In general,industries have moved away from operational expenditure(opex)software spend during every downturn.Instead,they have taken advantage of the situation to invest in more long-term solutions,such as software capital expenditure(capex).For more details on spending during various economic

66、 downturns,see Appendix 1.2000201020202000201020202000201020202000201020202000201020202000201020202000201020200.02.04.00.00.51.06.00.00.51.01.50.000.250.500.750.00.51.01.50.00.10.20.30.40.00.20.40.60.8%SoftwareDevicesR&DCommunication servicesIT servicesInternet&cloudAdvertisingSource:Oxford Economic

67、s estimatesSECTION 3:TECH SPEND AS A DRIVER OF BUSINESS OUTCOMES1415Section 3:Tech spend as a driver of business outcomesKEY FINDINGS AND TAKEAWAYS From correlation analysis,we find that sectors that spend heavily on tech appear to experience greater growth,especially in the long term.Keeping tech s

68、pend a consistently high share of the total thus seems crucial for success as well as resilience.During Covid-19,industries in the middle of the pack averaged much larger revenue losses compared to those at the top.We verify these observations via econometric analysis using a dynamic panel model and

69、 controlling for the influence of other sources of revenue.We find that maintaining or even raising opex is particularly important for supporting revenue growth in downturns.The largest and most significant impact is found for the Covid-19 downturn.Our estimates suggest that many industries in manuf

70、acturing and the public and private services sectors underspent in tech opex during the GFC downturn and overspent in tech capex during the Covid-19 pandemic.MORE THAN JUST A STRONG CORRELATION BETWEEN TECH SPEND AND REVENUE GROWTH IN NORMAL TIMES AND DURING DOWNTURNSOur analysis reveals a nuanced,m

71、ultifaceted connection between technology adoption and revenue growth across different industries.While a positive correlation is evident(as shown in Chart 6),our econometric analysis uncovers a deeper dynamic(see main results section for details).We observe a virtuous cycle where accelerated revenu

72、e growth fuels technology investment,which in turn amplifies business performance.This effect is particularly strong when comparing companies within the same industry group,especially those dedicating a larger share of costs to technology.Chart 6:Tech spend as a driver of revenue growth5012345Heavym

73、anufacturingLightmanufacturingPrivateservicesPublicservicesTradesTransportationand storageAvg.revenue growth(1997-2021),inflation adjusted(%)HighLowTech maturity5 High tech maturity includes industry-country observations with an above median level of teach spend relative to revenues.Source:Oxford Ec

74、onomics estimates16Section 3:Tech spend as a driver of business outcomesOur review of the literature points to several factors driving this phenomenon.Enhanced operational efficiency,technologys role in unlocking new revenue streams,and the stimulation of innovation all play a role.Since 1997,indust

75、ries with consistent,high-level technology spending have not only seen efficiency gains but also strategically positioned themselves to capitalize on emerging opportunities in the long run.Notably,this trend held true even during economic downturns,with industries choosing to cut back more on techno

76、logy experiencing disproportionately greater revenue declines.Overall,sustained technology investments demonstrate a dual benefit:enhanced efficiency and proactive market engagement in good times,but also resilience and adaptability in challenging economic climates.TECH SPEND HAS BEEN IMPORTANT FOR

77、RESILIENCE AND RECOVERYSimple correlations also suggest that tech spend has played an important role for resilience and recovery.Those industries that had a relatively higher level of tech spend as a share of revenues before the recession saw a milder impact in terms of revenue growth.And,despite st

78、arting from a higher base,industries with higher tech spend also saw stronger recovery growth rates in the two years that followed the GFC recession.On average,post-downturn revenue growth was between 0.4 and 2.2 percentage points higher in industries with above median tech spend as a share of reven

79、ues.BOX 2:A REVIEW OF THE DYNAMIC PANEL MODEL USED IN THIS STUDYA dynamic panel model(DPM)is a type of regression model that has become the standard tool economists use to understand the effects of both time and individual characteristics on a depen-dent variable.In this case,a DPM is ideal to direc

80、tly estimate how revenue growth has been influenced by changes in technology expenditures,its own past values,as well as by other explanatory variables,such as the size of the industry and exposure to large-scale economic downturns(i.e.,global recessions).However,estimating a dynamic panel model acc

81、urately is not straightforward because it involves some technical challenges.One of them is the problem of endogeneity,which means that some of the explanatory variables may be correlated with what is left“unexplained”by the regression equation.This can bias the estimates of the coefficients and lea

82、d to wrong conclusions.Several steps were taken to address this problem:First,the variables included in the regressions were transformed to remove the common influence of time trends.In addition,the dependent variable and select regressors were expressed in relative terms to ensure comparability acr

83、oss periods and countries(e.g.,tech investment as a share of industry revenues).Additionally,lagged variable values were included as instruments because these were found to be correlated with the explanatory variables,but not with the“error term.”These instruments were assumed to be predetermined an

84、d exogenous and were used to project the endogenous variables in the regression equation to provide consistent estimates of the coefficients.One challenge that arises when using lagged values as instruments is that lagged values may not be sufficiently correlated with the current values of the expla

85、natory variables,especially if there is a lot of variation over time or across industries.This can weaken the power of the instruments and reduce the precision of the estimates(a problem typically known as“weak instruments”).To overcome this problem,we used a systems-generalized method of moments(GM

86、M)approach,proposed by Blundell and Bond(1998),which combines two sets of equations:one in levels and one in first-differences.The first-difference equation eliminates the individual effects that may be correlated with the explanatory variables,while the level equation preserves some information tha

87、t may be lost in differencing.By combining these two equations,more instruments can be used,and this also increases their validity and strength.1617Section 3:Tech spend as a driver of business outcomes1 OPERATING EXPENDITURES IN TECH HAVE BEEN IMPORTANT DRIVERS OF REVENUE GROWTHAnalytical results fr

88、om our econometric study suggest that,in general,industries with higher operating expenditures in tech are more likely to experience stronger growth.The analysis showed that tech opex had a positive and significant impact on annual revenue growth when controlling for other factors specific to each i

89、ndustry.For every 10%increase in tech opex over a three-year period,the annual revenue growth rate increased by an average of 0.150.2 ppts.This was a substantial return on investment,given that tech opex typically accounted for 2%to 3%of revenues.Moreover,tech opex also had a delayed effect on reven

90、ue growth(Chart 7),implying that underinvesting in this area could have a cumulative negative effect on revenues.Chart 7:Estimated impact of a 10%increase in tech opex over three years on revenue growthImpact duringfirst 3 yearsAdditional impact onthe following 3 yearsPercentage points(annual avg.)0

91、0.050.100.150.20Our results show that the impact is larger when such expenditures remained relatively high as a share of revenues.Expenditures that were sustained for over two years were also found to have much higher influence on future revenue growth.This aligns well with OECD analysis(2015),which

92、 claims that higher-tech industries experience a larger benefit from R&D investments.Our findings indicate that tech capital expenditures(capex)have not played a significant role in bolstering revenue growth,contrary to the prevalent belief that tech capex is more crucial than operational expenditur

93、es(opex),particularly during crises.This suggests that many leaders may have adopted a less effective strategy by disproportionately investing in tech assets rather than focusing on tech operations.For instance,tech operations encompass aspects like cloud services,cybersecurity measures,and software

94、 subscriptions,which directly contribute to the agility and efficiency of a business.These operational tools have shown a direct impact on customer engagement,data management,and overall business adaptability.In contrast,tech assets,such as heavy investments in proprietary data centers or expensive

95、hardware,while important,may not immediately contribute to revenue growth.In some cases,these assets can lead to underutilized resources and increased fixed costs without a proportional return on investment.By highlighting these examples,we aim to make a clearer distinction between the types of tech

96、nology investments that can directly support revenue growth and those that,though valuable,might not offer the same immediate impact.Data deep-dive reveals four key takeaways for decision-makersSource:Oxford Economics estimates18Section 3:Tech spend as a driver of business outcomes2 OPEX TECH EXPEND

97、ITURES BECOME EVEN MORE IMPORTANT WHEN THE INDUSTRY FACES A RECESSIONTechnology spending is crucial for firms to gain a competitive advantage,increase efficiency,and foster innovation.These expenditures are even more important when an industry faces a recession,as they can help it survive and recove

98、r faster.On the flipside,this implies that industries with a lower share of tech expenditures are more vulnerable to weaker growth.The impact of tech opex is on average 4x higher during these periods compared to normal times.The relationship was particularly strong during the Covid-19 downturn but h

99、olds across all downturns included in the 25-year sample(Chart 1).This finding is aligned with a range of studies and indicators also pointing toward the importance of spending decisions during an economic downturn6.Tech capital expenditures were not found to have a larger impact during recessions i

100、n the immediate term.However,our results do provide empirical evidence that together opex and capex can play an essential role leading up to a recession from a resilience perspective.Looking at the 5-to 6-year period after the 2008 global financial crisis,we find that higher tech spend leading up to

101、 and during this recession generally supported post-recovery growth.Again,these results suggest a larger and more direct influence of the share of opex spend on post-downturn growth relative to capex expenditures.Still,businesses that invested more in tech capex during the crisis had stronger post-r

102、ecovery growth rates than those that did not.These results are consistent with two 2020 McKinsey studies7:One that reviewed more than 500 firms,and found that those that increased their tech spending by at least 10%during recessions had an average 8%higher revenue growth and 6%higher profit margin i

103、n the following three years,compared to those that cut their tech spending by more than 10%.Another that revealed organizations that made more capital expenditures in digital technology during the crisis were twice as likely to report outsize revenue growth compared to other companies.6 See evidence

104、 presented in Appendix 1 on Accenture(2008),McKinsey(2018),Microsoft(2020),Johnson Controls(2023)7 See McKinsey studies on Covid-19 and the tech tipping point(2020).Businesses that invested more in tech capex during the crisis had stronger post-recovery growth rates than those that did not.19Section

105、 3:Tech spend as a driver of business outcomes3 IN DOWNTURNS,BUSINESSES HAVE PRIORITIZED TECH CAPEX DESPITE HIGHER IN-RECESSION RETURNS IN OPEXChart 8 provides a detailed look at the downturn trends of tech opex and capex during the last two recessions across different industries.Overall,the respons

106、e of tech spend to downturns has been mixed.Chart 8:Tech spend and revenue growth in downturnsWithin the goods-producing sectors,businesses in heavy manufacturing have tended to maintain their relative level of tech opex unchanged while raising the level of capex as a share of revenues.In the light

107、manufacturing industries,tech opex was highly cyclical before Covid-19.Businesses in transportation and storage made large tech capex cuts after the GFC but held up relatively well during the recession that followed.Tech spend in the other services sectors,including private services and the trades,e

108、xperienced relatively large declines in response to the GFC.After the pandemic,there were broad increases in tech spend across sectors,but this was overwhelmingly concentrated in capex,especially in manufacturing and private and public services.Given the industry-specific regression results reported

109、 in the appendix,one big takeaway is that many businesses in manufacturing and the public and private services industries underspent on tech opex during the GFC downturn and overspent on tech capex during the Covid-19 pandemic.Our regression results also show that,generally,tech capex spend plays a

110、larger role for supporting resilience in transportation and storage and private services compared to other sectors.When it comes to tech opex,cross-sectoral differences are less evident,though resilience in retail and wholesale trade seems to be less dependent on tech compared to other sectors.%(inf

111、lation adjusted)-100-20-100-20%(inflation adjusted)GFCCOVIDGFCCOVIDGFCCOVIDGFCCOVIDGFCCOVIDGFCCOVIDHeavy manufacturingLight manufacturingPrivate servicesPublic servicesTradesTransportation and storageCapExOpExRevenuesSource:Oxford Economics estimates20Section 3:Tech spend as a driver of business out

112、comes8 Spiezia,V(2012).ICT investments and productivity:Measuring the contribution of ICTS to growth.OECD Journal of Economic Studies,Vol.2012/1.Available at http:/dx.doi.org/10.1787/eco_studies-2012-5k8xdhj4tv0t9 Lee,H et al(2016).Impact of IT Investment on Firm Performance Based on Technology IT A

113、rchitecture.Procedia Computer Science,Volume 91,2016,Pages 652-661 Available at https:/doi.org/10.1016/j.procs.2016.07.1644 OF THE DIFFERENT TYPES OF TECH SPEND,ENTERPRISE SOFTWARE AND R&D SPEND SEEM TO PLAY A CRUCIAL ROLE SUPPORTING RESILIENCE AND GROWTH,BUT THERE IS GREAT VARIATION ACROSS SECTORS.

114、We also explored the aggregate impact of different subcomponents of tech spending within operating and capital expenditures.Within opex,software,advertising,internet and cloud services,and R&D appear to provide more support for growth in general,while spending on devices,IT services,and communicatio

115、n services plays a less influential role.During downturns,opex spending cuts in software and R&D are relatively more harmful for resilience compared to the other categories.Within capex,software spend is again the most influential,but devices and IT services also play an important role.During downtu

116、rns,capex spending cuts in devices and IT services are relatively more harmful for resilience compared to other capex categories,but relatively less compared to in-downturn opex cuts.Academic literature has only been partially aligned on the distinctive impact of different technology products.For ex

117、ample,Spiezia(2012)finds that computing hardware gives the biggest boost to growth8,whereas Lee et al(2016)proclaim wireless technologies are the one crucial category to invest in for revenue growth9.While few studies make the granular distinction between products that this paper is presenting,two i

118、nteresting facts emerge from literature:Newer studies tend to find a higher importance of software and cloud-related products,while older studies tend to find a bigger impact of hardware and communications.Looking at more recent time series vs.older time series,the impact of tech spend on growth and

119、 profitability seems to be increasing over timesuggesting that technologies may be getting more advanced and productive,but also humans understand them better and are more skillful in employing them.Newer studies tend to find a higher importance of software and cloud-related products,while older stu

120、dies tend to find a bigger impact of hardware and communications.21SECTION 4:CONCLUSION2122Section 4:ConclusionTo summarize,this report offers a valuable perspective on how technology investments affect industry performance in times of economic crisis.THREE FINAL TAKEAWAYS ARE WORTH HIGHLIGHTING:1 K

121、eep up with the latest technologies It is important for business leaders to maintain or increase their tech spending in line with their industry standards,especially in times of crisis and uncertainty.Our analysis shows that consistent tech spending has a positive effect on growth over time and can

122、help businesses gain a competitive edge and ensure long-term survival.Software,R&D,and internet and cloud services are particularly crucial for resilience,as they enable productivity and innovation.Business leaders should be careful not to reduce their spending in these areas.2 Optimize your tech sp

123、end mixWhile our findings suggest that technology investments can boost resilience,competitiveness,and readiness for difficult situations,business leaders should also be strategic and prudent.They should avoid making hasty or risky technology choices based on hype or pressure.Interestingly,tech opex

124、 had a higher return on investment than tech capex during the last two recessions,but industries tended to invest more in tech capex than opex.This implies that many industries could have improved their tech spend mix to use technology more efficiently and effectively in times of trouble.3 Understan

125、d your business environment and macroeconomic context To help business leaders make informed decisions for their specific domains,we have provided detailed appendices at the end of this document that offer in-depth analyses of tech spend in different parent sectors.Executives are advised to compare

126、these industry-level insights with firm-level data to discover new patterns,identify opportunities and challenges,and adopt best practices that characterize each industry.SECTION 5:AREAS FOR FURTHER INVESTIGATION2324Section 5:Areas for further investigationWhile our research addresses several key qu

127、estions about how tech spending affects performance across different industries during and after recessions,other related questions could not be assessed due to limitations in data availability and quality.We highlight some of the issues below,which we hope to address in future qualitative and quant

128、itative research.1)What specific types of tech spending adjustments(such as shifting to cloud,investing in automation,or enhancing cybersecurity)were more effective for improving post-recession productivity growth during Covid-19,and will likely be key going forward?The seven tech categories in this

129、 study offer a lot of insight,but some additional(and very specific)sub-categories have become relevant for business success in the current environment,such as generative AI or green computing.Therefore,augmenting the dataset to include more differentiation within the tech spend categories by type w

130、ill be important to understand how these emerging technologies matter across different industry groups.Fetching this sort of data likely requires survey methods and/or quantitative estimation.2)How did the impact of tech spend on revenues and resilience vary across different countries,granular indus

131、tries,and company sizes and what did that do to market share?More availability of business-level data(instead of industry aggregates)would help us explore how tech spending impacts resilience and growth across countries.Vu(2004)10 offers some insight here,arguing that the difference in tech spend im

132、pacts could be explained by a combination of openness,education,institutional quality,and language skills meaning that regions like the US and Western Europe reap higher benefits from tech spend than emerging and developing markets.However,existing studies on this topic are dated,and revisiting thes

133、e with an additional 20 years of data could yield fresh insights on cross-country differences in tech spend impacts.In addition to exploring cross-country industry differences at a granular level,a dataset with firm-level data would be useful in examining how the results differ across various types

134、of enterprises.These datasets are typically collected through survey methods.Looking at existing literature,findings suggest larger enterprises enjoy bigger impacts of tech spend vs smaller firms,11 which may be due to economies of scale,but those studies tend to focus on a single country only,rathe

135、r than comparing across countries as in our study.3)How did the companies that maintained or increased their overall tech spending compare to those that reduced or stopped it after the last recession and in previous recessions?There are significant lags in reporting in the data sources that report t

136、he granular level of data needed to update the tech spend dataset.As more data on the post-Covid period becomes available we hope to update this review and estimate the Covid-specific and post-recovery dynamics more accurately.10 See Vu,K.(2004).Measuring the Impact of ICT Investments on Economic Gr

137、owth.11 See OECD(2015),Feng(2018)and Chhaidar(2022).25Section 5:Areas for further investigation4)Should the post-Covid surge in tech valuations be attributed to tech hype,and did this cause a subsequent hangover?Or has there been a visible shift to a new era of tech?An updated version of this datase

138、t will be necessary to produce reliable projections of tech spending by industry and type,as our data series ends in 2021 and therefore does not capture recent events.Dynamics observed in the stock market and tech-sector layoffs reported in the news can provide an early indication that there was ind

139、eed a hangover stemming from unrealistic expectations that were formed during the early months of the pandemic in 2020,in part comparable to the early 2000s.However,as both workplace and consumer behavior have normalized to some extent over the last year (i.e.,returned to pre-Covid standards),tech s

140、pend has not followed those optimistic trends.Nevertheless,tech spend may remain permanently higher given the increasing speed of technological innovations and the increasing importance of tech adoption for competing in the market.Moreover,our analysis suggests many firms still underspend on tech,so

141、 there is the potential for additional“catch up”spending that was only partially realized during Covid-19.Conversely,it is important to keep in mind that other factors remain important for business success,such as securing talent(and,in the case of producing firms,sourcing inputs).Thus,the expectati

142、on of large and rapid upward shifts in tech spend are likely overoptimistic.Overall then,we suspect the recent correction and job losses in the tech sector was simply a correction driven by the macroeconomic environment rather than a harbinger of decline for the tech sector,and that fundamentals sti

143、ll point to the increasing importance of the industry in the long run.1212 The recent volatility in tech jobs is likely the result of over-hiring during pandemic followed by layoffs in 2022-23.While the layoff numbers are significant at tech companies,especially Google,AWS,etc.,many have rosters tha

144、t are still much larger than pre-pandemic in absolute terms.And a Bureau of Labor Statistics report(2018)highlights that the high-tech sector tends to recuperate jobs lost during downturns relatively quickly.2526Section 5:Areas for further investigation5)What factor does human capital play in realiz

145、ing the productive capabilities of tech?How has spend on human capital interacted with spend on technology?Adding analysis on the importance of labor in driving technology adoption and effectiveness and,indeed,on the potential displacement of labor through capital was,although an important aspect,no

146、t within scope of the present study.Combining our results with broader literature findings suggests a strong link between human and technological capital,though.For instance,the work of Vu(2004)and Okere(2018)13 suggest that human capital is a key driver of the effectiveness of technology investment

147、,especially as firms become more globalized and exposed to a competitive market.Cathles(2020)14 verifies these results and suggests labor and digital technologies are complementary in driving firms financial performance.These are important findings and applicable for setting strategic business prior

148、ities especially as we enter the beginnings of a new technological era centered around artificial intelligence.Based on empirical and anecdotal evidence surveyed,firms should keep in mind the following:When companies start tightening their belts,employees are often the first to be trimmed.But axing

149、employees can be more costly in the long term than retaining them given their complementary with technology and contribute to a reduction in corporate stakeholder value.Organizations that invest in the right highly skilled technology talent will operate more efficiently,weather economic downturns be

150、tter,recover more quickly,and emerge in a stronger competitive position.Recent findings from academic literature confirm that the main factor behind resilience during downturns is innovation supported by organizational agility,a concept encompassing,among other factors,digitization,skills and sustai

151、nability practices.15 13 Okere,H.(2018).The impact of investment in technology and human capital on the firms financial performance:An empirical study.International Journal of Business Research and Information Technology,vol.5,no.1,winter 2018,pp.79+.Available at https:/ 14 Cathles,A.;Nayyar,G.and R

152、ckert,D(2020).Digital technologies and firm performance:Evidence from Europe.EIB Working Papers,No.2020/06.Available at https:/doi.org/10.2867/3688815 In their study,Bughin et al(2021)examine the performance of large global firms during the Covid-19 pandemic.Survey data was collected on 4,100 firms

153、around the globe,with questions focused on in-house views on innovation,organizational agility,and digitization and sustainability practices.They examine firm performance and find four statistically significant drivers of corporate resilience:disruptive innovation,agility in business structure,accel

154、erated digitization and sustainability practices,and engagement in platform-based business ecosystems.Combining our results with broader literature findings suggests a strong link between human and technological capital 27Appendix 1:Tech spend during recessions 28Appendix 2:Industry-specific analysi

155、s 31Appendix 3:About our econometric analysis 37APPENDICES28Appendix 1:Tech spend during recessionsAPPENDIX 1:TECH SPEND DURING RECESSIONSDuring economic downturns,the types of tech spend cuts have varied significantly after each of the past three recessions.For example,global devices spend was one

156、of the hardest-hit types of tech spend after the global financial crisis(GFC).However,it remained resilient during 2020 as companies shifted their focus to supporting the infrastructure needs of people working from home.In general,industries have moved away from operational expenditure(opex)software

157、 spend during every downturn.Instead,they have taken advantage of the situation to invest in more long-term solutions,such as software capital expenditure(capex).Tech spend growth during recessionsSoftware OpExSoftware CapExR&D OpExR&D CapExIT Services OpExIT Services CapExInternet&Cloud OpExDevices

158、 OpExDevices CapExCommunication services OpExAdvertising OpEx-20%-10%0%10%202020092002Source:Oxford Economics estimates29Appendix 1:Tech spend during recessionsOverall,the approach to corporate technology investment in the aftermath of both the 2008 global financial crisis and Covid-19 downturns con

159、trasted sharply with that of post-dotcom,both based on ideological views and economic realities.During the dotcom bubble,there was a period of excessive expectations and hype around the American technology sector,driven by the rapidly unfolding era of online business and digital services.When the bu

160、bble burst,austerity measures prevailed.Only a quarter of technology budgets were allocated to new IT initiatives,with the remaining funds going toward capital expenditures for maintenance.16 Executives during the dotcom crisis viewed IT with skepticism and cut their technology budgets to survive th

161、e downturn,17 but by 2008 executives were beginning to understand that technology was no longer a nice-to-have.Still,given the severity of the 2008 recession,the IT industrya useful stand-in for corporate technology spendingexperienced a greater decline than it did during the dotcom crisis.But spend

162、ing was likely more compressed by forced deleveraging than by management choice.18 Based on our analysis of tech spend as a share of revenues,cuts may have gone too far and slowed down revenue growth below what would have been possible under a more optimal technology investment strategy.During and a

163、fter the GFC,tech executives spoke out against what they considered outmoded views on spending during downturns and started to reinvigorate the push for more tech investments to achieve faster growth,resilience,and future competitiveness.19 And,indeed,companies that continued to strategically invest

164、 in the eras core technologies and build out their infrastructure were better positioned for survival and Tech spend in last two downturns200620072008200920102011201720182019202020212022YearAvg.across industries,share of revenues(%)GFCCovid-194.64.85.05.25.45.05.45.86.26.6Avg.across industries,share

165、 of revenues(%)16 See article on Computer World(2001):Gartner:Anemic IT spending growth projected for 2002.17 In sharp contrast with todays attitudes,Decision Economics Inc.s chief global economist at the time remarked in the LA Times that in a bad economy,“nothing is indispensable,not even new tech

166、nology.”Shortly afterwards,Nicholas G.Carr warned in his Harvard Business Review article“IT Doesnt Matter,”that overspending on IT is one of the greatest risk companies face,highlighting the sentiment toward tech spend during the dotcom crisis.18 See Kanerva,M.,&Hollanders,H.(2009):The Impact of the

167、 Economic Crisis on Innovation and OECD(2012):Science,Technology and Industry Outlook 2012.19 In March 2008,Accentures former Chief Technology Strategist Bob Suh wrote an opinion piece in Computerworld warning executives that viewing technology as a cost rather than an investment would be a huge mis

168、take going into the financial crisis.Source:Oxford Economics estimatesSource:Oxford Economics estimates30Appendix 1:Tech spend during recessionsgrowth in the years to comeas evidenced,for instance,by a 2018 McKinsey study on recession and post-recession performance of a large group of companies.20 T

169、he 2008 recession established that digital acceleration could separate companies that merely survived from those that grew after an economic downturn,but the amount of IT investment and the speed of digital transformation seen during the Covid-19 pandemic were unprecedented.In the beginning of the C

170、ovid-19 pandemic,it quickly became clear that early adopters of cloud computing and other innovative technologies were better equipped to handle the disruptions that followed worldwide lockdowns.This idea was adopted in both the tech-producing and tech-purchasing industries,highlighted by the follow

171、ing examples:Julia White,corporate vice president of Microsoft Azure,pointed out that companies with more advanced e-commerce platforms or AI-enabled supply chain forecasting tools outperformed those that lagged in these areas.21 Microsoft as well as the broader tech sector,of course,have a strong i

172、nterest in this storyline being true,but the development of our technology spend dataset makes it possible to verify these assertions empirically.Building-systems management company Johnson Controls,for example,does not expect to cut its digital projects despite concerns of a looming recession.“We h

173、ave to continue to invest,”says the companys chief technology officer Vijay Sankaran,who is steadfast in his digital strategy.“You cant do what were trying to do without technology.”22 The need for businesses to stay connected to consumers and workers resulted in a massive ramp-up in investments int

174、o digitizing infrastructure.As more granular industry data becomes available in the coming years,we will be able to produce precise estimates of the extent to which this increase has reversed in 2022-23 or whether the pandemic pushed industries to permanently spend more on tech on a relative basis.2

175、0 McKinsey tracked and analyzed 1,000 large,publicly traded companies across industries and geographies between 2007 and 2011,identifying a group of companies,labeled resilients,that outperformed their industrys median rate of total returns to shareholders(TRS)growth during the recession and after.2

176、1 See Duffy,S.(2020):One of Big Techs top moneymakers is getting a pandemic boost.CNN Business.Available at:https:/ See Lohr,S.(2023):Beyond Silicon Valley,Spending on Technology Is Resilient.The New York Times.Available at:https:/ have to continue to invest.You cant do what were trying to do withou

177、t technology.Vijay Sankaran Chief Technology Officer,Johnson Controls31Appendix 2:Industry-specific analysisAPPENDIX 2:INDUSTRY-SPECIFIC ANALYSISPRIVATE SERVICESAs a share of revenues,tech spend in private services has experienced gradual growth from 1997 to 2016.The increase was primarily driven by

178、 capex growth which outpaced opex in 2017,supported by strong growth in devices and IT services expenditures.Since the pandemic,opex tech spend growth has been concentrated in internet and cloud services and advertising,while the surge in tech capex has been supported by strong R&D,devices,and softw

179、are spend.Results from our empirical analysis suggest that,in normal times,advertising,software,and internet and cloud services spend play a larger role supporting revenue growth.During recessions,tech opex spend is more important relative to capex,but no one category is significantly more important

180、 than the others.These results are consistent with the possibility that this industry may have overspent in software capex in response to the Covid-19 pandemic.Tech OpEx and CapEx:Private servicesTech spend evolution by type:Private services(percent of revenues)CapExOpEx024620002005201020152020Perce

181、nt of revenues(%)2000201020202000201020202000201020202000201020202000201020202000201020202000201020200.02.04.00.00.51.06.00.00.51.01.50.000.250.500.750.00.51.01.50.00.10.20.30.40.00.20.40.60.8SoftwareDevicesR&DCommunication servicesIT servicesInternet&cloudAdvertising%Source:Oxford Economics estimat

182、esSource:Oxford Economics estimates32Appendix 2:Industry-specific analysisPUBLIC SERVICESAfter a period of steady growth after the dotcom bubble,tech spend in public services plateaued after the GFC recession primarily driven by relative declines in advertising,devices,and communication services.Whi

183、le internet and cloud and software spend continued to see rapid growth after 2008,they remained relatively small compared to the other categories.Tech spend in this industry saw a relatively less dramatic rise in response to the Covid-19 pandemic.The increases in R&D,advertising,internet and cloud s

184、ervices,and software were largely offset by declines in devices and communication services.Tech spend is less correlated with revenue growth for this industry in normal times.However,during recessions,opex in software appears to have been an important driver of resilience.With the exception of commu

185、nication services,results show that in-recession spend in the other categories may have also played a positive albeit more modest role supporting revenue growth,especially IT services and internet and cloud.There is no clear evidence that businesses in this sector either over or under spent in tech

186、during the pandemic.Tech OpEx and CapEx:Public servicesTech spend evolution by type:Public services(percent of revenues)CapExOpEx0.01.02.020002005201020152020Percent of revenues(%)0.51.52000201020202000201020202000201020202000201020202000201020202000201020202000201020200.00.10.00.20.40.20.00.40.80.0

187、0.20.40.60.00.51.00.00.10.20.30.00.20.40.60.3SoftwareDevicesR&DCommunication servicesIT servicesInternet&cloudAdvertising%Source:Oxford Economics estimatesSource:Oxford Economics estimates33Appendix 2:Industry-specific analysisTRADESAfter some modest increases at the beginning of the millennium,tech

188、 spend growth in businesses in retail and wholesale trade has been relatively weak.As a share of revenues,tech opex had been on the decline after 2008,driven by a continued drop in devices and communication services spend.Tech capex remained relatively stable.During the Covid-19 pandemic,increases i

189、n software,internet and cloud,and advertising opex were offset by declines in communication services which dropped considerably during the lockdowns.Regression results suggest that in-recession spend in devices,internet and cloud services,and advertising is most important from a resilience perspecti

190、ve.Importantly,these results do suggest that businesses in this industry may be underspending in devices,especially during downturns.There is no evidence that these industries overspent in tech in response to Covid-19.Tech OpEx and CapEx:TradesTech spend evolution by type:Trades(percent of revenues)

191、CapExOpEx0.02.04.020002005201020152020Percent of revenues(%)1.03.02000201020202000201020202000201020202000201020202000201020202000201020202000201020200.000.040.00.20.40.080.00.20.30.00.51.00.00.51.00.00.10.20.30.00.20.40.60.120.1SoftwareDevicesR&DCommunication servicesIT servicesInternet&cloudAdvert

192、ising%Source:Oxford Economics estimatesSource:Oxford Economics estimates34Appendix 2:Industry-specific analysisTRANSPORTATION AND STORAGEPre-Covid,businesses in the transportation and storage industries had kept a stable level of tech spend as a share of revenues over almost two decades.Interestingl

193、y,tech capex dropped after the dotcom bubble,while opex remained resilient.Then,in response to the GFC it was tech capex that remained stable while there were some declines in opex primarily driven by cuts in devices and communication services.In contrast to most other industries,opex tech spend out

194、grew capex in response to Covid-19.This was primarily supported by sharp increases in internet and cloud services,advertising,communication services,and software.The econometric analysis shows that,for businesses in this industry,revenue growth is highly influenced by changes in R&D expenditure.Devi

195、ces,communication services,and internet and cloud services also play an important role.In-recession capex spend was estimated to support resilience indicating that this industry may have under invested during Covid-19.At the same time,negative regression results suggest businesses may have overspent

196、 in communication services in response to the pandemic.Tech OpEx and CapEx:Transportation and storageTech spend evolution by type:Transportation and storage(percent of revenues)CapExOpEx0.01.02.020002005201020152020Percent of revenues(%)0.51.5200020102020200020102020200020102020200020102020200020102

197、0202000201020202000201020200.00.10.00.20.40.20.00.20.00.20.60.00.51.00.00.10.20.00.20.40.10.4SoftwareDevicesR&DCommunication servicesIT servicesInternet&cloudAdvertising%Source:Oxford Economics estimatesSource:Oxford Economics estimates35Appendix 2:Industry-specific analysisHEAVY MANUFACTURINGBusine

198、sses in the heavy manufacturing industries held a relatively low level of tech spend as a share of revenues in the two decades prior to Covid-19.Both capex and opex dropped after the dotcom bubble and in response to the GFC but with a one-year lag.Surprisingly,tech spend actually increased in 2019 e

199、ven as the global manufacturing sector fell into recession.Like most other industries,capex tech spend outgrew opex in response to Covid-19.This was primarily supported by sharp increases in software,devices,and internet and cloud services.Regression results suggest increasing IT services,software,a

200、nd communication spend can have a larger impact on revenue growth during normal times.Despite being relatively large,R&D spend is also an important driver,but the returns on investment are relatively lower.Software capex spend is the most important during recessions,followed by devices and R&D.There

201、 are no clear signs of businesses in this industry overspending in tech during Covid-19,but there could have been additional benefits from higher opex expenditures primarily in internet and cloud services.Tech OpEx and CapEx:Heavy manufacturingTech spend evolution by type:Heavy manufacturing(percent

202、 of revenues)2.5CapExOpEx0.01.02.020002005201020152020Percent of revenues(%)0.51.52000201020202000201020202000201020202000201020202000201020202000201020202000201020200.000.030.00.20.40.090.01.50.00.10.30.00.20.60.000.060.090.00.10.21.00.20.060.50.40.030.30.62.00.12SoftwareDevicesR&DCommunication ser

203、vicesIT servicesInternet&cloudAdvertising%Source:Oxford Economics estimatesSource:Oxford Economics estimates36Appendix 2:Industry-specific analysisLIGHT MANUFACTURING Both tech capex and opex in light manufacturing remained resilient after the dotcom bubble.However,capex jumped in response to the GF

204、C while opex entered a period of gradual decline.As in heavy manufacturing,tech spend actually increased in 2019 even as the global manufacturing sector fell into recession,but this was primarily driven by higher investments in tech capex.Capex tech spend also outgrew opex in response to Covid-19 su

205、pported by sharp increases in software,devices,and R&D spend.Regression results suggest R&D and software are the most important tech spend categories supporting revenue growth during normal times.However,businesses in this industry seem to have overspent in tech capex during Covid-19 including R&D.T

206、ech OpEx and CapEx:Light manufacturing Tech spend evolution by type:Light manufacturing(percent of revenues)CapExOpEx0.01.02.020002005201020152020Percent of revenues(%)0.51.52000201020202000201020202000201020202000201020202000201020202000201020202000201020200.000.00.51.00.20020.00.10.30.00.20.60.000

207、.100.150.00.20.20.1010.40.050.41.53SoftwareDevicesR&DCommunication servicesIT servicesInternet&cloudAdvertising%Source:Oxford Economics estimatesSource:Oxford Economics estimates37Appendix 3:About our econometric analysisAPPENDIX 3:ABOUT OUR ECONOMETRIC ANALYSISWhile the correlations observed in the

208、 data suggest a positive relationship between investments in technology and companies growth prospects,Oxford Economics undertook a more sophisticated econometric analysis to gain deeper insights into the impacts of various types of tech spend on growth and resilience,especially during economic slow

209、downs and downturns.Specifically,a dynamic panel model approach was used to analyze the survey-based dataset of industry-level technology expenditures described in Section 1 to better identify the causal relationship between tech spend and revenue growth.As explained in Box 2,this methodology for re

210、gression analysis was selected because it offers a rigorous framework to estimate and compare the effects of technology expenditures across different sectors and countries and how they change over time.Importantly,systems-generalized method of moments(GMM)regressions were performed directly on the l

211、arge dataset,taking advantage of its richness to generate more comprehensive insights.This approach also allowed us to control biases stemming from the complex dynamic relationships between revenue growth and other drivers reflected in past performance including productivity,innovation,competition,a

212、nd regulation.Our modeling confirms that maintaining tech spend during a recession does not just raise in-downturn resilience,but also likely boosts post-recession growth.For businesses,this means the right mix of tech spend during economic downturns has far-reaching implicationswith the potential t

213、o drive gains in market share relative to tech spend laggards.Similar to in-recession outcomes,we find evidence that making adjustments to opex(rather than capex)has a stronger and more direct impact on post-downturn growth.This means companies should proceed with particular caution when deciding on

214、 potential tech-related opex cutbacks during recessions.38About Oxford EconomicsABOUT OXFORD ECONOMICSOxford Economics was founded in 1981 as a commercial venture with Oxford Universitys business college to provide economic forecasting and modeling to UK companies and financial institutions expandin

215、g abroad.Since then,we have become one of the worlds foremost independent global advisory firms,providing reports,forecasts and analytical tools on more than 200 countries,100 industries sectors,and 8,000 cities and regions.Our best-in-class global economic and industry models and analytical tools g

216、ive us an unparalleled ability to forecast external market trends and assess their economic,social and business impact.Headquartered in Oxford,England,with regional centers in New York,London,Frankfurt,and Singapore,Oxford Economics has offices across the globe in Belfast,Boston,Cape Town,Chicago,Du

217、bai,Dublin,Hong Kong,Los Angeles,Mexico City,Milan,Paris,Philadelphia,Stockholm,Sydney,Tokyo,and Toronto.We employ 600 staff,including more than 350 professional economists,industry experts,and business editorsone of the largest teams of macroeconomists and thought leadership specialists.Our global

218、team is highly skilled in a full range of research techniques and thought leadership capabilities from econometric modeling,scenario framing,and economic impact analysis to market surveys,case studies,expert panels,and web analytics.Oxford Economics is a key adviser to corporate,financial and govern

219、ment decision-makers and thought leaders.Our worldwide client base now comprises over 2,500 international organizations,including leading multinational companies and financial institutions;key government bodies and trade associations;and top universities,consultancies,and think tanks.September 2024A

220、ll data shown in tables and charts are Oxford Economics own data,except where otherwise stated and cited in footnotes,and are copyright Oxford Economics Ltd.The modeling and results presented here are based on information provided by third parties,upon which Oxford Economics has relied in producing

221、its report and forecasts in good faith.Any subsequent revision or update of those data will affect the assessments and projections shown.To discuss the report further please contact:David Schockenhoff Head of US Macro Consulting Oxford Economics|USA Sebastian Herrador-Guzman Lead Economist Macro Con

222、sulting Oxford Economics|Canada This publication contains information in summary form and is therefore intended for general guidance only.It is not intended to be a substitute for detailed research or the exercise of professional judgment.Member firms of the global EY organization cannot accept resp

223、onsibility for loss to any person relying on this article.3940AppendicesGlobal headquartersOxford Economics Ltd Abbey House 121 St Aldates Oxford,OX1 1HBUKTel:+44(0)1865 268900London4 MillbankLondon,SW1P 3JA UKTel:+44(0)203 910 8000FrankfurtMarienstr.15 60329 Frankfurt am MainGermanyTel:+49 69 96 75

224、8 658New York5 Hanover Square,8th Floor New York,NY 10004USATel:+1(646)786 1879Singapore6 Battery Road#38-05Singapore 049909Tel:+65 6850 0110Europe,Middle Eastand Africa OxfordLondonBelfastDublinFrankfurtParisMilanStockholmCape TownDubaiAmericasNew YorkPhiladelphiaBostonChicagoLos Angeles TorontoMexico CityAsia PacificSingaporeHong KongTokyoSydneyEmail:Website:Further contact details: