Robotics, Automation and the Economy Rob Seamans May 19, 216
Why Do We Care? Automation Productivity growth Bloom, Sadun, Van Reenen (AER 212) Industrial robotics are type of automation, so Robots (?)Growth Graetz & Michaels (215):.36 to annual productivity growth Industrial robotics can complement or substitute for labor At the industry level and at the occupation level Timely: recent popular press interest, July 215 JEP on automation, ongoing NAS study (Brynjolfsson and Mitchell) 216 Economic Report of the President, Chapter 5 (Technology and Innovation) 1
Automotive Assembly Line 2
Automated Warehouse Systems 3
Increase in Worldwide Robotics Shipments Estimated Worldwide Annual Shipments of Industrial Robots Thousands of Units 24 225 21 179 18 165 159 15 12 112 114 12 113 12 97 9 6 6 3 24 25 26 27 28 29 21 211 212 213 214 Source: International Federation for Robotics, World Robotics 215 The Boston Consulting Group (214) has estimated that worldwide spending on robotics will be $26.9B in 215 and will rise to $66.9B by 225. Annual industrial robotics shipments have nearly doubled since 21. 4
Automotive Leads All Industries Thousands of Units 7 6 5 4 3 Estimated Annual Supply of Industrial Robots by Main Industries: 21-212 21 211 212 Robot Density: Automotive vs. Non-automotive (212) Number of Robots per 1, Workers 1,8 1562 1,5 1,2 1133 191 9 Automotive Average Non-automotive 2 1 Automotive Electrical & Electronics Chemical, Rubber, and Plastics Source: International Federation for Robotics, World Robotics 214 Metal Food 6 332 273 3 219 147 135 76 213 2 11 Japan Germany United States China Source: Axium Solutions; International Federation for Robotics Automotive industry leads other industries, in number and number-per-worker. Increase since 21 seems pronounced in automotive. Japan leads Germany and U.S. 5
Link between Occupational Prob. of Automation and Wages Podiatrists, civil engineers, clergy interior designers, HR managers Median Wages and Probability of Automation by Occupation Median Hourly Wage (21) 9 8 7 6 5 4 3 2 1.2.4.6.8 1 Probability of Automation Source: Bureau of Labor Statistics; Frey and Osborne (213); CEA Calculations Tax preparers, telemarketers, cashiers, office clerks Subjective, forward-looking measure (Frey and Osborne 213) Occupations that are easier to automate have lower wages. E.g., low reliance on manual dexterity, originality/creativity, social perceptiveness, negotiation and persuasion skills. 6
Link between Occupational Prob. of Automation and Wages Podiatrists, civil engineers, clergy interior designers, HR managers Median Wages and Probability of Automation by Occupation Median Hourly Wage (21) 9 8 7 6 5 4 3 2 1 Economists.2.4.6.8 1 Probability of Automation Source: Bureau of Labor Statistics; Frey and Osborne (213); CEA Calculations Tax preparers, telemarketers, cashiers, office clerks Subjective, forward-looking measure (Frey and Osborne 213) Occupations that are easier to automate have lower wages. E.g., low reliance on manual dexterity, originality/creativity, social perceptiveness, negotiation and persuasion skills. 7
Upstream Activity: Recent Uptick in Patenting Activity Number of Patents 4 35 3 Patents with Robot Class, 2 214 Share of All Patents (right axis) Number of Patents (left axis) Percent of All Patents.13 214.12.11 25 2 15 1 5.1.9.8.7.6.5 2 22 24 26 28 21 212 214 Source: United States Patent and Trademark Office. Patenting activity has started to increase since 212, both in number and in rate. No evidence of concentrated ownership across industries. 8
Summary So Far 1 What We Know Robotics sector is small ($27B vs. $2T for mfg). Recent uptick since 21 (units, revenues, patents). Upstream supply seems competitive. Potential for productivity growth, but also potential for labor displacement. 2 What We Hope to Know Need a better understanding of when robots (and automation) are substitute vs. complement. Need to better characterize the impact by geography. More research on effect on productivity and growth 9 3
Next Steps Researchers: continue to work with existing data IFR shipments (country-industry-year level) Subjective assessments of probability of automation Patent applications w/ robot class Census data on ICT, e-business, etc. Researchers: develop (and share) new data sources Systematic U.S. survey e.g. Last year, how much money did your establishment spend on robotics? and Has your establishment considered using robotics instead of human labor? Policymakers: track affected industries and geographies NSTC subcommittee on AI/ML 1