EF: Another branch of your research has focused on how management practices affect firm and country productivity. Why do you think management practices are so important?In the New Zealand context its interesting to note Bloom's comments on the effects of foreign direct investment. Multinational firms help improve the standard of management in the host country and thus help improve productivity. Add this to the point that Eric Crampton noted about foreign firms paying their employees more, then foreign investment in New Zealand looks better than the anti-FDI crowd would have us believe. Also Bloom highlights the advantages of a flexible labout market. Given the small size of the internal New Zealand market Bloom's point about the importance of competition to good management practices emphasises the need to keep New Zealand open to trade so that local produces face as much competition as possible from foreign firms.
Bloom: My personal interest was formed by working at McKinsey, the management consulting firm. I was there for about a year and a half, working in the London office for industrial and retail clients.
There's also a lot of suggestive evidence that management matters. For example, Lucia Foster, John Haltiwanger, and Chad Syverson found using census data that there are enormous differences in performance across firms, even within very narrow industry classifications. In the United Kingdom years ago, there was this line of biscuit factories — cookie factories, to Americans — that were owned by the same company in different countries. Their productivity variation was enormous, with these differences being attributed to variations in management. If you look at key macro papers like Robert Lucas' 1978 "span of control" model or Marc Melitz's 2003 Econometrica paper, they also talk about productivity differences, often linking this with management.
Economists have, in fact, long argued that management matters. Francis Walker, a founder and the first president of the American Economic Association, ran the 1870 U.S. census and then wrote an article in the first year of the Quarterly Journal of Economics, "The Source of Business Profits." He argued that management was the biggest driver of the huge differences in business performance that he observed across literally thousands of firms.
Almost 150 years later, work looking at manufacturing plants shows a massive variation in business performance; the 90th percentile plant now has twice the total factor productivity of the 10th percentile plant. Similarly, there are massive spreads across countries — for example, U.S. productivity is about five times that of India.
Despite the early attention on management by Francis Walker, the topic dropped down a bit in economics, I think because "management" became a bad word in the field. Early on I used to joke that when I turned up at seminars people would see the "M-word" in the seminar title and their view of my IQ was instantly minus 20. Then they'd hear the British accent, and I'd get 15 back. People thought management was quack doctor research — all pulp-fiction business books sold in airports.
Management matters, obviously, for economic growth — if we could rapidly improve management practices, we would quickly end the current growth slowdown. It also matters for public services. For example, schools that regularly evaluate their teachers, provide feedback on best practices, and use data to spot and help struggling students have dramatically better educational outcomes. Likewise, hospitals that evaluate nurses and doctors to provide feedback and training, address struggling employees, and reward high performers provide dramatically better patient care. I teach my Stanford students a case study from Virginia Mason, the famous Seattle hospital that put in place a huge lean-management overhaul and saw a dramatic improvement in health care outcomes, including lower mortality rates. So if I get sick, I definitely want to be treated at a well-managed hospital.
EF: How much of the productivity differences that you just discussed are driven by management?
Bloom: Research from the World Management Survey that Raffaella Sadun, John Van Reenen, and I developed suggests that management accounts for about 25 percent of the productivity differences between firms in the United States. This is a huge number; to give you a benchmark, IT or R&D appears to account for maybe 10 percent to 20 percent of the productivity spread based on firm and census data. So management seems more important even than technology or innovation for explaining variations in firm performance.
Coincidentally, you do the same exercise across countries and it's also about 25 percent. The share is actually higher between the United States and Europe, where it's more like a third, and it's lower between the United States and developed countries, where it's more like 10 to 15 percent.
Now, you may not be surprised to learn that there are significant productivity differences between India and the United States. But you look at somewhere like the United Kingdom, and it's amazing: Its productivity is about 75 percent of America's. The United Kingdom is a very similar country in terms of education, competition levels, and many other things. So what causes the gap? It is a real struggle to explain what it is beyond, frankly, management.
EF: What can policy do to improve management practices?
Bloom: I think policy matters a lot. We highlight five policies. One is competition. I think the key driver of America's management leadership has been its big, open, and competitive markets. If Sam Walton had been based in Italy or in India, he would have five stores by now, probably called "Sam Walton's Family Market." Each one would have been managed by one of his sons or sons-in-law. Whereas in America, Walmart now has thousands of stores, run by professional nonfamily managers. This expansion of Walmart has improved retail productivity across the country. Competition generates a lot of diversity through rapid entry and exit, and the winners get big very fast, so best practices spread rapidly in competitive, well-functioning markets.
The second policy factor is rule of law, which allows well-managed firms to expand. Having visited India for the work with Benn Eifert, Aprajit Mahajan, David McKenzie, and John Roberts, I can say this: The absence of rule of law is a killer for good management. If you take a case to court in India, it takes 10 to 15 years to come to fruition. In most developing countries, the legal system is weak; it is hard to successfully prosecute employees who steal from you or customers who do not pay their invoices, leading firms to use family members as managers and supply only narrow groups of trusted customers. This makes it very hard to be well managed — if most firms have the son or grandson of the founder running the firm, working with the same customers as 20 years ago, then it shouldn't be surprising that productivity is low. These firms know that their sons are often not the best manager, but at least they will not rampantly steal from the firms.
The third policy factor is education, which is strongly correlated with management practices. Educated and numerate employees seem to more rapidly and effectively adopt efficient management practices.
The fourth policy factor is foreign direct investment, as multinational firms help to spread management best practices around the world. Multinational firms are typically incredibly well run, and that spills over. It's even true in America, where its car industry has benefited tremendously from Honda, Toyota, Mitsubishi, and Volkswagen. When these foreign car manufacturers first came to America, they achieved far higher levels of productivity than domestic U.S. firms, which forced the American car manufacturers to improve to survive.
The fifth factor is labor regulation, which allows firms to adopt strong management practices unimpeded by government. In places like France, you can’t fire underperformers, and as a result, it's very hard to enforce proper management.
EF: Do you expect America's productivity advantage to continue?
Bloom: On the above five criteria, the United States scores an "A" on four of them except education, where we score a "C." The United States has a weak school system and poor education standards compared to a number of our competitors. For example, based on OECD Pisa [Programme for International Student Assessment] scores, the U.S. educational system ranks in the mid-20s on math, below many European and East Asian countries. So improving educational standards is the most obvious way to improve management and ultimately growth, because poor education makes it harder to manage our firms. Fixing U.S. education will take more funding. But most importantly, it will require dismantling the cobweb of restrictions that teachers unions and politicians have put on schools, like tenure and seniority-based pay.
If you fix these five drivers of management, you're 95 percent of the way there. Most other factors seem of secondary importance compared to the big five of competition, rule of law, education, foreign direct investment, and regulations.
EF: Management practices can be viewed as "soft" technologies, compared to so-called "hard" technologies such as information technology. Do you see anything special about the invention and adoption of these "soft" technologies relative to "hard" technologies?
Bloom: The only distinction is that hard technologies, like my Apple iPhone, are protected by patents, whereas process innovations are protected by secrecy.
The late Zvi Griliches, a famous Harvard economist, broke it down into two groups: process and product innovations. Most people who think of innovation think of product innovations like the shiny new iPhone or new drugs. But actually a lot of it is process innovations, which are largely management practices.
Good examples would be Frederick Winslow Taylor and scientific management 100 years ago, or Alfred Sloan, who turned a struggling General Motors into the world's biggest company. Sloan pushed power and decision-making down to lower-level individuals and gave them incentives — called the M-form firm. It seems perfectly standard now, but back then firms were very hierarchical, almost Soviet-style. And then there was modern human resources from the 1960s onward — the idea that you want to measure people, promote them, and give them rewards. Most recently, we have had "lean manufacturing," pioneered by Toyota from the 1990s onward, which is now spreading to health care and retail. This focused on data collection and continuous improvement.
These have been major milestones in management technologies, and they've changed the way people have thought. They were clearly identified innovations, and I don't think there's a single patent among them. These management innovations are a big deal, and they spread right across the economy.
In fact, there's a management technology frontier that's continuously moving forward, and the United States is pretty much at the front with firms like Walmart, GE, McDonald's, and Starbucks. And then behind the frontier there are a bunch of laggards with inferior management practices. In America, these are typically smaller, family-run firms.
EF: What are the key challenges for future research on management?
Bloom: One challenge is measurement. We want to improve our measurement of management, which is narrow and noisy.
The second challenge is identification and quantification: finding out what causes what and its magnitude. For example, can we quantify the causal impact of better rule of law on management? I get asked by institutions like the World Bank and national governments which policies have the most impact on management practices and what size impact this would be? All I can do is give the five-factor list I've relayed here; it's very hard to give any ordering, and there are definitely no dollar signs on them. I would love to be able to say that spending $100 million on a modern court system will deliver $X million in extra output per year.
One way to get around this — the way macroeconomists got around it — is to gather great data going back 50 years and then exploit random shocks to isolate causation. This is what we are trying to do with the World Management Survey. The other way is a bit more deliberate: to run field experiments by talking with specific firms across countries.
Wednesday, 24 December 2014
Management and productivity: An interview with Nicholas Bloom
The following comes from an interview with Stanford University economist Nicholas Bloom available online at "Econ Focus", the economics magazine of the Federal Reserve Bank of Richmond.