GLU: The True Cost of Doing Business

There is a belief widely shared among policymakers that if arguments for a proposal or decision are supported by numbers on a page then somehow this makes that choice less political.

By Conor Cradden, originally published in the Global Labour Column

There is a belief widely shared among policymakers that if arguments for a proposal or decision are supported by numbers on a page then somehow this makes that choice less political. It permits the claim that what is being proposed is not really a choice at all but something that the ‘evidence’ demands. This emphasis on quantitative indicators has meant that much policy argument has been displaced into the design of the indicators themselves. Rather than being grounded on purely technical criteria, the design of statistical indicators is a highly politicized process in which different stakeholders struggle to ensure the numbers that emerge will be more compatible with arguments in favour of their policy predilections than those of the opposition.

The World Bank’s ‘Doing Business’ (DB) indicators are a shining example of statistics that come with this kind of built-in value judgment. The DB indicators claim to be a guide to the relative ease of establishing and running a business in different countries. This is ‘measured’ on a number of dimensions, including starting up, paying taxes, getting construction permits and enforcing contracts. The indicators allow the construction of rankings, including an overall global ranking that places Singapore at the top – making it the world’s easiest place to do business – and Chad at the bottom.

This might appear to be an innocent enough endeavour. While states obviously have the right to ensure that there is a proper measure of social and political oversight of economic activity, it is also obvious that oversight procedures can be more complicated and more expensive than necessary. However, although the Bank denies that the DB indicators encourage deregulation, the information the indicators provide gives no way of judging whether the cost of conforming with regulation is reasonable in the light of the social, economic and environmental benefits that it produces. They have nothing to say about whether a country might on the whole be better off because of regulation. Since the social costs associated with deregulation are invisible to the DB indicators, governments whose concern is to improve their position in the DB ranking – and in some cases this is even a condition of financial aid from the Bank – have no incentive to take the potentially negative effects of deregulation into account.

Nowhere is the assumption that regulation is only a cost clearer than in the case of the ‘employing workers’ (EW) sub-indicator. A country’s EW score depends on the cost of making employees redundant and a measure called ‘rigidity of employment’, which is a composite index where the highest possible score corresponds with a low minimum wage for beginning employees, easy availability of fixed-term rather than permanent contracts, minimal restrictions on night and weekend working, high maximum permitted weekly working time, a low number of days of paid holiday and minimal requirements for notice and consultation when making redundancies.

Not surprisingly, the EW indicator has attracted criticism from many directions, but most notably the global labour movement. The ICFTU criticised the DB indicators within weeks of their first publication in 2003. Since then the Confederation, and subsequently the ITUC, has set out objections on a number of occasions, both in direct communication with the Bank and in public papers. In 2007, the ILO joined the debate, producing an official paper[2] that criticised the EW indicator on technical grounds, but also because of what it called problems with ‘policy coherence’ – in other words, the EW indicators cut directly across the ILO’s own, arguably more legitimate policies. The ILO argued that the view that “reducing protection to a minimum and maximizing flexibility is always the best option” was badly mistaken and that the EW indicator was “a poor indicator of the investment climate and labour market performance”.

The paper sparked a series of exchanges between the ILO and the Bank that culminated in the establishment of a consultative group (CG) to serve as a ‘source of advice’ on revising the EW indicator. Around the same time – early in 2009 – pressure from the global unions led to the Bank agreeing that at least until the group reported, the EW indicator would not be included in the calculation of the overall DB ranking nor used as a basis for policy advice. The consultative group included senior Bank and ILO officials together with global union, employer and OECD representatives. There were also three independent members, a labour law expert, a social entrepreneur and a public servant.

The ILO’s decision to participate in the CG will not have been taken lightly – even though in principle all of the members were acting in their personal capacity. Not participating would have meant missing a rare opportunity to have an impact on an influential indicator, but participating was arguably a gamble. The risk was that the group would come up with conclusions that did not adequately respond to the ILO’s criticisms but that the Bank would put its recommendations into effect anyway. If the ILO wanted to object, it would be forced to get into a public argument with the Bank about the adequacy of an indicator in whose revision two of its senior officials had just participated.
Now that the CG has produced its final report[3] it is not obvious that the gamble paid off. The solution proposed to the principal problem – the fact that lower standards of labour protection receive a higher score – is hardly adequate. Three elements of the indicator – minimum weekly rest periods, paid holiday entitlement and the level and means of setting the minimum wage – have been changed from being in a simple inverse relationship with the indicator score (the lower the better) to a kind of ‘banding’ system in which the policy target is to have these protections fall within a lower and an upper limit. Not enough holiday and a country will not receive the maximum possible score, but the same is true for what is deemed to be too much holiday. A similar change is proposed for maximum weekly working time. The ranking on the minimum wage indicator for countries that have one remains inversely related to the ratio of the wage to the average value added per worker, but countries that have no minimum wage no longer receive the best possible score. This is reserved for systems in which the minimum wage is set by collective bargaining – as long as it applies to less than half the manufacturing sector, or does not apply to firms not party to it – and systems in which trainees or apprentices are excluded.

The report of the CG makes it clear that it was split on whether the changes to the EW indicator are adequate. ‘One view’ was that the modifications dealt with the substantial problems and that the EW indicator should be reintegrated into the overall DB indicators. A ‘second view’, on the other hand, “noted that EWI did not adequately reflect worker protections even after the amendments made, and that the Doing Business report should reflect labour regulations holistically, or not at all”. This second view also argued that if the EW indicator was to continue to be used, there should also be a separate, quantitative ‘worker protection measures’ indicator published alongside the DB indicators. However, although this idea was discussed by the CG[4], it failed to agree a recommendation on the issue.

The ILO now has to decide whether to carry on working with the Bank. If it does not, the Bank will probably put the modified indicator back into use, and may also go back to basing policy advice on the EW indicator. Certainly the ILO doesn’t have to endorse the revised indicator, but if it wants to avoid a public argument, the best it can do is maintain a studied neutrality on the issue. The fact remains, though, that the DB indicator is still a barrier to the improvement of working conditions and quietly accepting its existence would be cowardly at best. The obvious question is why the ILO does not try to take the collaboration implied in the consultative group one step further and to work to persuade the Bank that there ought indeed to be an official, jointly developed worker protection indicator. The stakes are not so high here since the ILO clearly has moral and technical authority on the issue that the Bank cannot claim.

So why the deafening silence from the ILO? There has been no comment on the report of the CG, still less any indication of whether the ILO wants to carry on working with the Bank. In fact, the problem for the ILO is less with the outside world than its own constituents. The possibility of producing a ‘decent work’ indicator has been floating around for more than 10 years. That such an indicator has not (yet) been developed is partly a reflection of the traditional reluctance of employers and governments to allow themselves to be ranked, and partly a reflection of disagreement about whether such an indicator should be focused on outcome measures – the extent to which decent work is a reality for workers on the ground – or regulation – the extent to which the formal rules conform with ILO policies. These are difficult questions, but making a determined effort to resolve them is likely to be less costly for the ILO than allowing the Bank to continue to use and promote its EW indicator.

[1] Disclosure: the author is married to an ILO official. The official in question has no input into ILO policy-making in the areas under discussion in this article.




Conor Cradden is a research fellow in the Department of Sociology at the University of Geneva and a partner in Public World, a London-based research and policy consultancy.