There’s a catch-22 problem with the management of industrial policy.
On the one hand, however rigorous the groundwork, all the requirements from an industrial policy instrument cannot be conceived ex-ante but will only show up in the course of its implementation. It’s therefore important to reliably capture the emerging feedback from the implementation. This requires close ties, even embeddedness, between the bureaucracy and business.
Based on the emerging feedback, it’s required to tweak, often significantly revise, the original policy framework. Such changes are generally likely to involve revisions of not only milestones and timelines, but also rates, slabs, amounts, instruments, and products. Since it’s hard to get data that can conclusively substantiate the changes, these decisions come with a lot of subjectivity and are essentially judgment calls. And such tweaking must often be done continuously over the years. This comes with a strong risk of capture.
It’s therefore required for the bureaucracy to exhibit some level of autonomy. And autonomy must be accompanied by accountability.
Effective industrial policy therefore demands what Peter Evans has described in the context of South Korea, embedded autonomy, where there’s also a balance between embeddedness and autonomy.
Réka Juhász, Nathan Lane, and Dani Rodrik described how the US Advanced Research Projects Agency (ARPA) operates.
The central figure in the ARPA model is the program director, who is not a career government official, but a professional from academia or industry who is recruited for a three-year term. The program director works in a designated area and selects a portfolio of projects to support with grants. The selection process and the associated performance milestones are designed following intense consultations and workshops involving the program director and the relevant stakeholders of researchers and firms. The program director works closely with the grantees as the projects unfold, reviewing and revising targets as needed.
Grantees are required to provide quarterly progress reports which the agency staff rates according to a traffic light system: projects that miss a critical milestone and are at risk of failing receive a “red”; projects that miss a milestone but can be expected to recover get a “yellow”; and projects that are on track to reach their goals get a “green.” Projects with red ratings call for more intense supervision and scrutiny on the part of program directors (in the form site visits, conference calls, meetings, and written analyses of problems and possible solutions). If putting the project back on track seems infeasible, the program director can issue a letter warning of the possibility of termination.
These performance criteria are coupled with considerable amount of flexibility. Specific milestones, like much else in these programs, are provisional: they are added or deleted in nearly half of the projects. Similarly, budgets can be expanded or decreased in light of project developments. In short, the model is based on continuous iterative collaboration with the private sector, and quite different from the hands-off, arms’-length, strict ex-ante-rules regulatory approach favored by economists.
Consider the example of an industrial policy scheme like production-linked incentives (PLIs) deployed in India to promote domestic manufacturing of medical equipment. The policy involves a list of products eligible for the PLI subsidy, the rates of subsidy, the timelines and milestones to avail of the subsidy, and the base on which the subsidy is calculated.
As the policy implementation proceeds, one comes across differential demand (or investment interest) for the products necessitating the need to double down on certain products, and conversely, also exit certain products. It also often emerges that the milestones and timelines must be continuously adjusted, and even the rate slabs must be tweaked due to unforeseen market-driven contingencies. Or it may be required to provide some complementary support (in terms of easing regulations; or facilitating access to some service like preferential treatment in public procurement or fast track clearances and permits; or convergence with some ongoing scheme) for certain products/equipment. Or allow specific firms the benefit of the doubt on certain eligibility conditions to access incentives.
Given that firms are making investment decisions in real-time, all the aforesaid policy decisions must be taken without going through the lengthy (and often interminable complete) process of changing the original policy guidelines. This necessitates having policy guidelines themselves flexible enough to give officials a very high level of discretion. It also requires a budget envelope that can be expanded, even if only at the margins, under certain circumstances to support some products.
Such discretion, in turn, necessitates strong governance systems. For a start, the selection, reporting, and monitoring systems should be as robust as mentioned in the case of ARPA. This should include strong mechanisms to discipline errant behaviours.
This should consist of guidance in terms of principles on when discretion can be exercised, evidence-based and well-reasoned exercise of such discretion, and administrative oversight of such decisions. This should be supplemented with a governance mechanism that’s sufficiently high-powered to make the aforesaid decisions and has the convening power to achieve effective coordination across different departments.
This kind of scheme iteration should be supplemented with a conscious choice to foster export competition and eliminate non-competitive firms.
Joe Studwell dwells at length on the ingredients that made the North East Asian economies successful in pursuing industrial policy. Two in particular stand out. The first one, of incentivising export competition, is widely known. But a second, perhaps more important requirement from political economy considerations and one which differentiated North East Asian economies from their less successful southern counterparts, was the focus on weeding out the failing firms (and promoting those succeeding).
In other words, create export competition among firms and let go of the weak. The policies and bureaucratic incentives were aligned to identify firms struggling to compete and turn off incentives to them, while promoting the competitive. The dynamics of competition when allowed to play out will leave the market with 2-3 big firms.
On the same lines, it’s therefore essential that going forward the PLI scheme incorporate provisions that explicitly seek out those firms that are failing and stop their incentives. Letting go of such firms is critical to aligning incentives across the ecosystem and improving the likelihood of success. Similarly, the policy must double down with support for the emerging bigger firms.
This will finally create a handful of big firms with global-scale contract manufacturing capabilities, and equally importantly, a big ecosystem of suppliers. At best, from these efforts could emerge one or two large branded manufacturers.
PS: The working paper has this nice snapshot on the dissonance between perception and reality on industrial policy!
President Reagan, famous for labelling government as the problem rather than the solution, protected U.S. steel, auto, and motorcycle industries from import competition to encourage them to invest in technological upgrading. President Pinochet of Chile, as close to an avatar of free market liberalism as one can imagine, subsidized the forestry sector and promoted its exports. Margaret Thatcher actively courted Japanese auto makers and promoted their UK investments with financial incentives
No comments:
Post a Comment