1. Introduction

This is a "concept" Regulatory Impact Assessment (RIA) indicating how and why there should be a comprehensive RIA before proceeding with the National Credit Bill. There is much confusion regarding RIAs because they are a new concept in South Africa. To bridge the gap, we summarise what RIAs are, and how to do them. More information is available from the Law Review Project (011-883-5843, [email protected]). To avoid needless repetition and length, we shorten such phrases as "Sub-section 130. (1) (c) (ii) (aa) of Part C of Chapter 6 of the Bill, if implemented in its present form as envisaged, is likely to have the following direct (primary) and indirect (secondary/unintentional) consequences or impacts", for example, to "The effect of §130 (1) (c) (aa) will be …".

1. Key Regulatory Impact Assessment (RIA) issues

Key issues, elaborated below, include:

1.1.1 RIAs have become or are becoming recognised internationally as a tool of good government in the form of a precondition for new laws and policies (jointly called "regulation").

1.1.2 RIAs enable governments to assess the probable effects of policies under consideration.

1.1.3 Specifically, their purpose is to quantify pre- and post-regulation conditions, and to compare them, so that governments know whether what they want to do is likely to have advantages ("benefits") exceeding disadvantages ("costs"), jointly called "impacts".

1.1.4 These impacts can, in turn, be compared with official objectives.

1.1.5 To that end, they are not, and should never be, policy or lobbying instruments. They help governments decide how to respond to policy proposals (including Bills) and lobbies.

1.1.6 Only accurate and comprehensive RIAs enable governments to appreciate the status quo and the probable effects of proposals – and to compare these with what they want to achieve – with sufficient certainty to justify proceeding.

1.1.7 Once expected impacts and policy objectives are known, the great secondary value of RIAs is that governments can (and should) monitor what happens in the real post-regulation world, so that they can improve or abandon measures not fulfilling their objectives.

1.1.8 RIAs should also include an assessment of whether all provisions are Constitutional, whether there is a realistic prospect of general compliance and enforcement, and whether powers and functions are in the appropriate Department.

2. Reasons for a "Concept" RIA

This is a "concept" RIA because it has not been possible to do a comprehensive assessment, for the following reasons:

Critical aspects of the proposed law are not in the Bill. Responsibility for these aspects is delegated to the Executive. This has two potentially serious implications:

1.2.1 Firstly, since it is not possible to know what substantive law will be made by the Executive (by way of regulation and policy), it is impossible to conduct a proper RIA. One of the requirements for good law is, in accordance with the separation of powers component of the rule of law (Section 1 of the Constitution), that its legal consequences are known to and adopted by the Legislature after compliance with elaborate checks and balances required by the Constitution and Parliamentary convention.

1.2.2 Secondly, the Bill empowers the Executive to make and change substantive law arbitrarily, which means the Legislature (Parliament and the DTI Portfolio Committee) cannot, in effect, know what law they are making, which means, in turn, they have no way of knowing whether benefits are likely to exceed costs.

1.2.3 Regarding those substantive provisions that are in the Bill, it was not possible in the time available or without the prospect of official recognition, to get all necessary inputs. Some of them require primary research and the estimates by diverse experts.

Where we have accurate figures, we use them; where not, we use illustrative numbers or indicate which numbers should be obtained. Experts may, without our knowledge, have access to some missing data. This RIA shows what conclusions can be drawn from known numbers, and what accurate numbers might indicate, were they available.

The most detailed assessments in this document relate to credit bureaux because the most readily available and most accurate data is provided by credit bureaux and the credit bureau Ombud.

We appreciate that there are experts who will recognise some of our illustrative numbers to be inaccurate. They may dismiss this document as uninformed. Apart from the reasons we have given for using inaccurate illustrative numbers (where we do not have accurate figures), we do not pretend to be experts on the credit market. We are experts on RIAs, and this is not a definitive RIA. It is a definitive illustration of the data a proper RIA would offer legislators, to enable them to make informed judgements. Where we have illustrative numbers, they can easily be replaced with accurate figures where or when those are available.

3. Principal findings

The most striking feature of our research was that:

1.3.1 Almost everyone consulted could provide quantified estimates of some costs – usually expecting them to be considerable – whereas virtually no one quantified reasons for the Bill – the mischief it seeks to redress – or expected benefits, except in general terms, such as "consumers will have better access to information." It was difficult to get anyone to predict how many would benefit, by how much, or by what process the Bill would achieve its objectives. Accordingly, there may not be an adequate rational connection between some provisions of the Bill and its objectives as required by the Constitution (the "rationality principle").

1.3.2 The Credit Law Review (CLR) documentation has virtually no quantification of the kind needed for a Regulatory Impact Assessment. Problems and benefits are asserted in general terms: there is "over-indebtedness", for instance, without concrete research on the nature and extent of the problem. There are no empirical estimates of the extent to which there will be fewer problems: how many people will benefit, in what form and to what extent. The CLR quantifies consumer perceptions and macro-aggregates (the amount and nature of credit in South Africa). It does not quantify the extent to which these consumer perceptions are objectively accurate, nor to what extent macro-aggregates are expected to change after the Bill is implemented.

1.3.3 On the available evidence, it is almost impossible to predict to what extent the objectives of the Bill will materialise. The evidence suggests that significant objectives will be achieved and others may be undermined. In the absence of evidence to the contrary, it must, for instance, be concluded that increased costs and risks for creditors will (a) reduce over-indebtedness and reckless lending, (b) increase the cost of credit, and (c) reduce access to credit, especially for less preferred consumers (SMMEs, the poor, historically disadvantaged people, people with negative records).

2. Introduction to Regulatory Impact Assessment (RIA)

This section is for people who want to know more about RIAs: what they are, what they are not, why government needs them, why one is especially important for this Bill, how to do them, and how to ensure they are objective. You can go straight to the RIA findings below.

1. World trends

There is a great deal of readily accessible information on RIAs on the Internet and in libraries, although it should be remembered that term "regulatory impact assessment" is the name for the concept in the United Kingdom (UK). RIAs go under different names in most other countries. They do not differ only in name. There are important differences in what they cover and how they are conducted.

This does not mean that we should duplicate the UK model uncritically in South Africa, but since it is one of the world’s most tried-and-tested, we can both learn much from it and benefit from the willingness of the British government to assist through its local representatives and its RIA practitioners in the UK. The Law Review Project (LRP) is preparing a detailed report, recommendations and guidelines for the government based on what it might draw from the world’s diverse experience with a view to producing the world’s newest and most appropriate RIA system for our circumstances.

Most mature first world democracies, and a rapidly growing number of regions (states, provinces, cantons and cities) within them, as well as developing countries, conduct pre-regulatory assessments that differ widely in name, institutions and form. Nevertheless, the objective is always the same: to enable governments to improve the quality and efficacy of their laws by identifying and quantifying relevant facts as accurately as can reasonably be done in advance. These summary notes are drawn from our study of the world’s RIA experience. They reflect the state of the art as it evolved over the years of trial and a great deal of error.

2. A tool for government

RIAs are the only way in which governments can assess the likelihood of law having desired effects, and establish in due course if they have done so. That makes it surprising that they took so long to become common practice in mature democracies, and why all legitimate governments have not adopted them. That RIAs are exclusively a tool of governance is not generally appreciated in South Africa. Some politicians and senior officials fear RIAs, not realising that RIAs serve their interests. Some private sector interests, on the other hand, believe mistakenly that RIAs can be used as a lobbying strategy.

As embryonic awareness of the fact that RIAs are under consideration spreads, all sorts of vested interests are becoming involved, and there is a danger of a good idea being hijacked or subverted by vested interests regarding it as a means to "clip government’s wings". All concerned have to appreciate that RIAs are not a lobbying instrument, and, when they are done for government, as in this instance, they have to be maximally objective aids to government decision-making, undertaken in good faith, preferably by recognised autonomous institutions or consultants with proven expertise. Submissions purporting to be RIAs have already been submitted to government. Those of which we are aware were prepared by people who were apparently unaware that there are recognised RIA methodologies. What they called RIAs were simply routine submissions representing narrow vested interests. They were called RIAs in a putative attempt to disguise them with a veneer of objectivity.

RIAs have been described succinctly as "a policy tool which assesses the impact, in terms of costs, benefits and risks, of any proposed law". They help policy-makers think through the probable positive, neutral and negative implications of proposals, including secondary or unintended consequences, thereby enabling governments to improve the likelihood of achieving their policy objectives.

3. Methodology

The RIA process, as opposed to the RIA itself, can be, if properly conceived and implemented, an important aid to good government. It not only improves the quality of advice that government obtains, but promotes more informed media and public opinion and discourse. More subtly, and perhaps more profoundly, RIAs protect governments from undue pressures and incentives to pass over-hasty, ill-considered and counter-productive laws, often in response to popular yet misinformed opinion (or perceived opinion). Governments are often subjected to a clamour for them to "do something" about real and imagined problems. This is often in the field of "social legislation", where laws promoting social objectives enjoy near-universal endorsement, such as child protection, public health and safety, or conservation. Where everyone agrees on the objectives of a law, governments are often denied the benefit of countervailing evidence, and may unwittingly adopt laws that are ineffective or impossible to implement at best, and excessively costly or counter-productive at worst. Governments are protected from these risks if RIAs are conducted properly. If the RIAs provide clear evidence that envisaged measures will probably result in costs exceeding benefits, as they often do, governments can enhance their efficacy and popularity by referring to RIAs as a reason for developing alternative and more promising solutions. For RIAs to serve this purpose, that is, for them to do what they were conceived to do, they must obviously be maximally accurate predictions of relevant factors (below).

4. Misuse of RIAs

Having summarised the objectives and considerable potential virtues of RIAs, a note of cautious realism is appropriate. The future is of course unknowable. RIAs cannot predict with certainty what will happen. As with forecasts generally, the future is unlikely to correspond precisely with what is predicted. If RIAs are conducted properly, they will give a much better indication of what is likely than poorly conducted RIAs and ones that are contrived to legitimise ill-conceived laws (to which we will return). Not only is there a huge difference between good (accurate) and bad (inaccurate) RIAs as predictors of an uncertain future, but bad ones are worse than none at all, because they give government and society false expectations, and can be politically embarrassing when promised results do not materialise.

Without RIAs, the world of government is somewhat unreal. Laws are made that determine the fate of nations; that condemn them to enduring misery or elevate them to paradigms of peace, stability and prosperity. Without RIAs, these laws are essentially made based on little more than the amorphous feeling that they are a good idea. There is nothing equating the kind of cost-benefit analysis taken for granted before business decisions are made. Specifically, there is little or no clarity on precisely why laws are needed at all (the "mischief principle"), what their direct and indirect costs (disadvantages) will be (for the department concerned, for other departments, for big or small businesses, for civil society, for rich and poor citizens), who will bear those costs, who will benefit, to what extent they will do so, and what secondary or unintended consequences there might be.

Not only is the future unknowable, but it is notoriously difficult to establish which causes have which effects in the social sciences. Even if minimum wage laws are followed, as they have been in South Africa (SA), by increased unemployment, and unemployment was identified in an RIA as a potential secondary effect, cause-and-effect cannot be concluded. The evidence would be considered in the light of what happened in other countries. Other factors can be considered, such as conspicuous changes in technology or currency values. Only when other possible determinants and the experience of comparable markets elsewhere have been considered, can there be confidence about the causality, and the need to repeal or revise the policy concerned.

5. How accurate and comprehensive?

There is huge range of accuracy and comprehensiveness. At one extreme, there are some that are no more than a page or two that amount to no more than an assertion that benefits will exceed costs. These are clearly a sham. At the other extreme, there are elaborate research projects with sophisticated assessments of probable effects of every provision. Mostly, they consider only core provisions, and rely on readily available data and probable effects estimated by experts. RIAs by people with vested interests for or against proposed measures tend to be predictably biased regardless of how elaborate they are.

Accuracy of data is often a problem, as it is regarding this RIA. In such cases, RIAs should indicate the plausible range, and the implications based on alternative assumptions. Often indicative or illustrative assumptions will suffice, provided they are reasonable and unbiased. What RIAs must do is give governments a good idea of what is likely to happen under alternative scenarios with sufficient accuracy and comprehensiveness for them to proceed with their choice in their chosen path.

6. Monitoring

Once expected benefits and costs, and secondary effects, have been quantified, government is able to monitor developments and see whether envisaged costs are exceeded, whether benefits materialise, and whether secondary effects are as benign as predicted. One of the innovations suggested for South Africa by the former head of the United States equivalent of the RIA Unit, the Office of Management and Budget (OMB), Wendy Gram, is that South Africa could be the first country to do RIAs on RIAs, to formally use RIA forecasts as benchmarks against their own accuracy and efficacy. Officials could be retained only if they have good records of accomplishment, and consultants could be remunerated in accordance with the accuracy of their predictions. Thus, they are rewarded on merit, and government ensures that its decisions are based on maximally accurate information. Furthermore, the government will have ongoing accurate information on which to base modifying the system to ensure a constantly improving RIA quality.

7. Dos and Don’ts

To minimise the risk to government of RIAs misleading them (to their subsequent embarrassment and the country’s cost) proven checks and balances are necessary:

2.7.1 RIAs must be consistent, complying with objective criteria. Such guidelines have been found to ensure consistency and reduce the prospect of RIAs being contrived to serve ends that were no part of the government’s intention. Consistency has many virtues, not least, that RIAs are presented to policy-makers in a form with which they are familiar.

2.7.2 The propensity to please. Politicians and senior officials are always subjected to the inclination of their staff and consultants to want to please them and to assume that they are expected to give the answers their superiors supposedly want. It must be clear to RIA practitioners from their guidelines that when the government is deciding what laws to make, it wants information that is maximally accurate. In short, it wants the facts.

2.7.3 To ensure accuracy and quality, RIAs should be undertaken by trained experts. If the experts are located in the government department proposing the law, they should be subject to audit (screening and approval) by an independent third party. An example is the British Regulatory Impact Assessment Unit in the Cabinet Office in the United Kingdom.

2.7.4 RIAs should be produced by autonomous agencies. The UK model achieves objectivity by having (a) special units within each government department that are responsible for producing RIAs, and (b) a special unit in the Cabinet Office which screens RIAs to ensure that they are not contrived by the relevant department to motivate or justify ill-conceived laws. Current thinking among experts in the field is that the approach increasingly adopted elsewhere is preferable, namely, to have RIAs themselves produced by autonomous agencies, which have no particular interest in the proposed measure, and are not subject to influence by the department concerned.

2.7.5 In South Africa, proposed laws are not subjected to mandatory screening to ensure jurisprudential quality. In other countries there is a tendency towards centralised drafting by specialised and autonomous drafters to further enhance the quality of laws. Jurisprudential and RIA functions can be combined fruitfully as they are in the aptly named Quality of Law Academy in the Department of Justice in The Netherlands. Governments presumably want to improve the quality of their laws in all senses. They want them to achieve desired consequences, to be properly drafted, to have benefits exceeding costs and to have them comply with the Constitution. To this end, the scope of RIAs can be broadened to include jurisprudential assessments. Accordingly, the RIA in this document assesses Constitutionality and jurisprudential quality.

2.7.6 RIAs can and should be done at different stages of the evolution of a law. The United Kingdom, for instance, has a four-stage approach. The first RIA is a simple assessment of whether the idea of the proposed law is worth pursuing at all. Following RIAs become increasingly sophisticated as the law’s detail is fleshed out. Although the RIA in this document is at the final stage of the Bill, a final RIA is not possible, because so much of the substance of the law is left to subsequent "subordinate" regulation.

2.7.7 An excellent innovation is for RIAs to consider alternatives. Instead of government being presented with a single option, they obtain an assessment of feasible alternatives. In the United Kingdom Prime Minister Tony Blaire made it mandatory for RIAs to compare the costs and benefits of the proposed law with (a) leaving the status quo, (b) repealing existing laws that might be causing the problem, and (c) at least one alternative solution if the first two were found wanting. There does not exist, and we have not produced, published alternatives with which to compare the Bill.

2.7.8 RIAs must not themselves be policy documents. They do not take a stand on measures under consideration. They merely say to and for government that "if you do this, that will happen". Strictly speaking, they are not therefore "cost-benefit" analyses, because they ideally make no assumption about what consequences are preferable. The RIA on a proposed dangerous dog registration law in the United Kingdom assessed the number and severity of dangerous dog incidents, the extent to which there would be less of these after the law, and the pecuniary and non-pecuniary cost per unit of reduction. It reported its findings blandly, without suggesting that costs would be "excessive" relative to benefits. Politicians decided that this was so, and the RIA enabled them to explain to the public’s satisfaction why they were not proceeding with a measure previously demanded by public opinion after a spate of much-publicised Pitbull Terrier attacks. They were aided by the "cooling-off" period necessitated by and the pubic awareness generated by the RIA process.

2.7.9 Impacts must be quantified. One of the principle objectives of RIAs is to "get beyond the clichés and platitudes of regulation", as the head of the Australian equivalent of the Britain's RIA Unit expressed it in an interview with a Law Review Project researcher. Before RIAs, laws would be justified on such nebulous grounds as being "in the public interest", "protecting consumers", "promoting orderly markets" and so on. RIAs establish what precisely will change. How many members of the public will benefit? To what extent? What will the unit cost be, in cash and kind, of each benefit? Regarding laws, such as this Bill, intended to protect consumers, what precisely will consumers be protected from? How many consumers will be protected? What is the value of the protection they will enjoy? At what cost will it be achieved? Who will bear that cost? How do these quantities compare with alternative solutions?

2.7.10 Everything is quantifiable. The most common obfuscation regarding RIAs – usually raised by narrow vested interests trying to conceal the fact that they expect generous benefits concentrated in the hands of a few at the dispersed expense of many (usually consumers or taxpayers in general) – is that the benefits of proposed laws or policies supposedly cannot be quantified, that they are "qualitative" not "quantitative". However, the RIA approach is that all impacts are quantifiable, and should be quantified. Quantification is obviously not necessarily monetary. Consider child protection and environmentalism laws: they are archetypal examples of measures with supposedly non-quantifiable "social" benefits. Yet, on reflection, it becomes clear that all important aspects are indeed quantifiable. How many children are abused? In what way? By whom? How often? What improvement will there be after the new law? For whom? How will the benefit be achieved? By whom? How much time, money and effort will be required?

It may be hard to quantify costs, especially secondary or unintended effects. Will improved consumer protection raise costs or reduce choices? Will it distort markets, discourage investors, create new incentives for corruption or abuse, or burden the courts unduly? If so, by what mechanism, to what extent, and at whose expense? The point about unintended consequences is precisely that they are hard or impossible to predict. However, they are potentially serious, if not fatal to the success of the measure, and one of the purposes of RIAs is to focus attention on identifying, anticipating and making best guesses regarding them.

3. Regulatory Impact Assessment on the National Credit Bill
1. General

1. There appears to be no hard evidence on whether the DTI’s sophisticated research on perceptions – which paints a bleak picture – is an accurate reflection of objective reality. If so, the government may want to improve reality; if not, the government may prefer to correct perceptions. The other main source of hard data published by the DTI is a study on The Cost, Volume and Allocation of Consumer Credit in South Africa by FEASIbilitY (Prof. Penelope Hawkins). We are unaware of benchmarks according to which costs, volumes, credit allocation and other variables in the study can be regarded as sub-optimal, which means there is no way of knowing whether impacts will improve matters. There are many other data sources regarding which we found no evidence of which quantified outcomes the government would prefer.
2. Some core provisions the law will create are not in the Bill. The Bill delegates power to the Executive to create far-reaching laws by regulation. Consequently, it is impossible for a RIA or the Portfolio Committee to assess important impacts.

Since it is impossible for the Portfolio Committee to assess probable costs and benefits, this concept RIA recommends that the Committee calls for:

3. an official RIA in which proposed and probable impacts are quantified, and
4. inclusion in the Bill of all substantive law.

It is impossible without a proper RIA for the Committee to:

5. predict whether the Bill’s benefits will exceed costs, or
6. even what laws the Bill will bring into effect.

• Will credit receivers be better off? If so, by virtue of what mechanism, in what way will they be better off and to what extent? Will they, for instance, have cheaper or more credit? If so, how much will they save or how much more credit will they get? Alternatively, will their benefit be that they get less credit, or less over-indebtedness? Will they have fewer judgments against them, and be more credit-worthy? Will they get higher rates of return on their investments? Or will they pay less on other people’s investments?
• Who will be worse off and to what extent? Will direct and indirect costs exceed benefits? Will compliance costs imposed on credit grantors exceed benefits to credit receivers? Will those who grant credit pass increased costs onto consumers and credit receivers, thus rendering them net losers?
• What will it cost government? Government will have to divert resources from elsewhere or increase taxes to implement and enforce the law. At whose ultimate expense will this occur and how much will it cost them?
• What will the secondary effects be? Regulations impact markets. What impacts will the law have? Will these impacts have net benefits or costs for the country and its economy? Will regulations such as standardisation and statutory consolidation curtail competition? If this is the case, what will be the cost and to whom?
• What about macro-economic policy? Do the recommendations support or undermine macro-economic policy? Should there be more or less credit in the economy? Should it be cheaper or should savers get better returns?
• Are the proposed powers consistent with the Constitution? Do they comply with other principles of good law? To what extent is the law enforceable? To what extent is it likely to be breached or ignored?
• When will impacts occur?
• Is there adequate provision for efficacy monitoring? What are the benchmarks? What corrective action will there be if expected costs are exceeded or benefits not achieved? Should all or some of the provisions be subject to sunset clauses?

1. Quantifiability

Whilst the impacts of all laws are quantifiable to a greater or lesser extent, impacts of financial laws, such as this Bill, are particularly quantifiable. It is possible, for instance, to calculate with uncommon precision how much more or less credit receivers obtain, in what forms they acquire it, and at what cost. Indeed, the Credit Law Review, which preceded this Bill, quantifies the cost and quantity of credit. It does not predict how the cost and quantity of credit will change. What else will change? Will there, for instance, be fewer judgements against consumers? Will consumers get less or more credit? Who will get less, and who will get more?

In section 2.6 of this document we mentioned the differing approaches in different countries, and within countries (at lower tiers of government). We do no go into the complex detail of the models from which we have chosen issues that are covered here.

We mentioned that RIAs in some countries describe at least two plausible alternatives, and give quantified reasons for rejecting them. The first is usually "Do Nothing", or retain the status quo. The second tends to be another regulatory option. The UK guidelines, for instance, require consideration to be given to repealing or relaxing existing laws that may be causing or exacerbating the problem. There should be detailed quantification of each option. We are in no position to do this here, but illustrate what form this section of the RIA should take.

2. Option 1 - Do Nothing

• There does not appear to be hard evidence of "market failure". There is a great deal of emotion and there are strongly held perceptions that the market is dysfunctional, but we find no "functional" benchmark against which the nature or extent of dysfunction becomes apparent. Indeed, the SA consumer credit market might be one of the best for a country at our stage of development, according to international ratings. Our credit information systems are, for instance, rated nearly as high as ratings go, and people with credit-worthiness ratings that would give them access to credit in comparable countries are accessing credit here. Our savings rate is low compared with high-growth developing countries, and a law making credit-granting more risky and costly might reduce the savings rate, on which the views of the Department of Finance and the Reserve Bank could fruitfully be solicited by the Portfolio Committee. One of the few pertinent research projects of which we know was conducted by the DTI as part of its Credit Bureau investigation. We do not know why the results were omitted from the Report. The Committee should demand that the results be presented to it, and published. In response to widespread criticism of credit bureaux, considerable cost and effort was invested in soliciting public complaints. These were investigated. The results were surprising. Only about 600 complaints were forthcoming. A random sample of 60 was taken for detailed investigation. A departmental representative called on the credit bureaux concerned and did not find a single justified complaint.
• There is general agreement that there are 15 million credit-active consumers. Of these, 80% have positive ratings. Both the absolute number of credit-active and credit-worthy consumers and the proportion with positive ratings are increasing at rates far in excess of economic growth (see below). Most newcomers are from historically disadvantaged groups. The status quo appears to be promoting the government’s objectives (as specified in the Long Title of the Bill) better than might reasonably have been expected.
• Between March and June 2005, there were 23,000 disputes with credit bureaux. This number is less than 0.128% of all credit-active consumers, which means one in nearly 800. Since there is an average of 2.6 credit agreements per credit-active consumer, the number of disputes is about 0.05% of credit agreements.
• On investigation, 87% of complaints were found by the credit bureaux to be unfounded, and corrections were made for the remaining 13%. Of the complaints considered by the credit bureaux to be unfounded, consumers referred 1250 to the Ombud. This means 0,007% (one in 14,400) were referred to the Ombud. The Ombud found 48% of the 1250 to be valid complaints, and ordered the bureaux to amend their records. That means that only 0.003% of credit-active consumers – one in 30,000 – had valid complaints that were not resolved expeditiously at minimal cost to all concerned by direct communication between them and the credit bureau concerned.
• What these figures suggest is that the laws governing credit bureaux (privacy, defamation, fraud, contract laws, etc.), and the Credit Bureau Code of Conduct, appear to create an appropriate legal environment for the effective functioning of the industry, at no significant cost to credit providers, consumers or the state.
• In contrast, the National Credit Bill entails considerable costs (below). If the Bill reduces legitimate complaints by an optimistic 50% – most experts expect it to be much less, if any – and if we assume a modest R20 million of the government’s proposed budget for this purpose, it will cost more than a disproportionate R34,000 per beneficiary.
• We could find no estimate of the damage suffered by consumers because of inaccuracies or irregularities. Damage suffered by consumers can be assumed to be above what they are willing to pay to have records corrected. Actual damage obviously varies enormously from case to case. It may average R100 per consumer. Even if we multiply this 10-fold, the average cost of R34,000 will far exceed the average benefit of R1000.
• Given these statistics, it is hard to imagine a scenario where benefits will exceed costs, at least as far as credit bureaux are concerned, because there appears to be very little scope for improvement, and the marginal cost of each unit of improvement seems likely to be excessive.
• Similar calculations on other intended benefits, such as free access to information, and the cost at which they would be achieved, yield similar pro status quo results. This suggests that serious consideration ought to be given to retaining the status quo as a viable alternative, which does not appear to have been done by any of the role players.
• In particular, a proper situation analysis is needed to establish whether there is a prima facie case for "doing something". That there is a great deal of agitation is presumably sufficient reason to investigate the factual nature and extent of problems beyond the rhetoric, and to consider response options, but probably insufficient for implementing extreme forms of intervention without thorough investigation of the kind for which the RIA approach has been developed.

1. Option 2 - Make industry-specific Codes of Conduct binding on all parties

All the affected industries have Codes of Conduct. Some of these are binding on the entire industry, including non-parties, by virtue of having been gazetted under the relevant law. These codes are usually supported by industry-funded Ombuds, and appear to function effectively. We gather from representatives of consumer organisations, trade unions and the industries concerned, that there is not much dissatisfaction with any of the Codes per se or their enforcement mechanisms. It appears as if attitudes range from near-universal satisfaction to the view that relatively minor provisions should be improved upon. The antipathy to some credit providers and credit bureaux appears to be more to do with scoring systems, incorrect records and real or suspected discrimination, rather than the substance of the Codes.

A comprehensive RIA would quantify costs and benefits of Codes of Conduct, differentiating them in each credit sector. We have not been able to do this but are assured that direct costs are minimal, and that benefits to debtors and competitors exceed costs. This means that the cost to the industry, and ultimately consumers, is so small when averaged through all transactions that it is inconsequential. The cost of processing an individual complaint may be a few thousand rand directly and indirectly, but in all sectors disputes subjected to dispute resolution proceedings are so few - always, seems, less than 1% - that the average cost per transaction, per credit-active consumer, or compared to the aggregate value of credit, is inconsequential.

Clearly, making Codes binding on non-parties in each sector and improving their provisions to the satisfaction of all parties will have much lower costs than the Bill. The question is whether the Bill would have benefits not achievable through such Codes, and whether whatever extra benefits there might be are sufficient to justify the extra enforcement, compliance and secondary costs implied by the Bill.

To establish this, we made extensive enquiries with a view to determining whether the Bill would have benefits beyond those of amended codes binding on non-parties. The consensus appears to be that there would be, but that they would be minimal. The principle additional benefit cited was that access to information presently entails costs to consumers whereas the Bill provides for free access. Here we make two observations:

3.4.1 Codes of Conduct could be amended to provided for all benefits under the Bill, in what is called "self-regulation". Whereas no attempt has been made to quantify differences, self-regulation is generally cheaper than statutory regulation in toto, and differs primarily in that costs are borne by the industry and its consumers rather than tax-payers in general. From an economic efficiency perspective, it is preferable for costs and benefits to contained within a sector.

3.4.2 "Free" access is never really free. Rights of access and error-correction contemplated by the Bill would have costs which will be paid by consumers indirectly instead of directly, because additional costs are passed onto consumers in general by way of higher costs-of-credit.

Whether "free" access is provided for in legislation or in industry codes, benefits accrue to individual consumers who take advantage of the right, at the expense of those who do not.

We cannot consider all differences between this option and the Bill here, so we consider only, by way of illustration, one of the most conspicuous differences, the establishment under the Bill of a National Credit Register. There is too little known about the numbers involved to make accurate assessments. Reasonable assumptions suggest that benefits added by the Register to what suitably amended Codes of Conduct can offer will probably be less than added costs. All costs of credit granting must, according to the laws of economics, be added to costs for and recovered from credit receivers. This is because, except in special circumstances, average rates of return on investments are determined by general economic forces and tend towards what economists know as dynamic "equilibrium". In other words rates of return fluctuate, but soon tend back towards the norm. Markets do, however, grow or contract according to whether impacts such as new laws drive suppliers out or attract new entry. A comprehensive RIA would assess such secondary impacts.

In the absence of quantified evidence to the contrary, it appears as if amending existing Codes to include benefits contemplated by the Bill, adding Codes for sectors without them – there do not appear to be any – and making them binding on non-parties, may have comparable benefits at lower costs, and have the added economic policy benefits of confining costs and benefits to the sector.

2. Option 3 - Repeal or reform existing laws and institutions

As explained above, this is an obligatory consideration under British practice, which is regarded as the world’s leading model. It is remarkably insightful because it acknowledges that existing problems for which regulatory solutions are proposed are often themselves a consequence of earlier regulation and, occasionally, deficiencies in common law, or inappropriate institutions within the government or the private sector.

There is an erroneous belief in South Africa to the effect that the industry is not already regulated. It is, of course, regulated by a multiplicity of laws and institutions, some of which are industry specific, and some of which are elaborate. A comprehensive RIA would identify all of these and consider whether they have net adverse effects (secondary or unintended consequences; costs exceeding benefits etc.). Only by analysing the status quo can a case be made against this option. To our knowledge, there has been no such analysis, and we have not been in a position to produce one.

In a global climate where almost every country and all major international agencies are concerned about over-regulation, the Portfolio Committee might consider it necessary to call for such an analysis before projecting this option.

3. Option 4 - The National Credit Bill

In RIAs in those countries where alternatives must be assessed, this section would commence with a description of why this option was selected in preference to the alternatives. The full report of the Technical Committee, whose recommendations informed the Bill, has not been published. We understand from a presentation made to us by Mr Gabriel Davel and from the summary, that the only alternative considered by the Technical Committee was whether there should be an interest cap (i.e. a usury limit). There is therefore no known comparison between alternative scenarios other than, as far as we know, the modest contribution of this RIA.

We now turn to the main substance of the Bill. Ideally, each substantive provision would be the subject of a quantified analysis of the kind illustrated above, in which the following questions are asked and answered quantitatively:

1. Future benefits - general

• Who will benefit?
• Profile of beneficiaries?
• Number of beneficiaries in specified categories?

• How will they benefit (what form will their benefits take)?
• To what extent will they benefit?

• How much more (or less) access to credit?
• How much more (or less) reckless lending?
• How much more (or less) over-indebtedness?
• How much more (or less) debt?
• How much more (or less) restructured debt?
• How much more (or less) access to information?
• How much less cost to consumers directly (and more cost indirectly)?

1. Future costs - general

• At whose expense will each benefit to consumers accrue?
• How much additional cost / burden?
• What indirect consequences will there be?

1. Existing costs and benefits - General

1. Lack of accurate data

• Fragmented and outdated legislation;
• Inefficient consumer protection, particularly in relation to consumers in low income groups;
• High cost of credit, and, in some areas, access to credit;
• Rising levels of over-indebtedness;
• Reckless behaviour by credit providers and exploitation of consumer by micro lenders, intermediaries, debt collectors and debt administrators."

1. Businesses who are protected from innovative and new competition. Big business often welcomes such measures because uniform "one-size-fits-all" regulation tends to reduce competition from more flexible smaller businesses
2. Officials in the private and government sectors for whom jobs will be created at the expense of alternatives from which resources are diverted
3. Politicians who capitalise on the apparent popularity of the measure
4. Consultants, some of whom are already offering compliance monitoring and training

1. Conflicting objectives

One of the problems that becomes immediately apparent in a RIA is that the Bill appears to have to conflicting objectives. It envisages a reduction of reckless lending and over-indebtedness and increased access. The evidence presented to us and, we understand from media coverage, the Portfolio committee, suggests that these two objectives may be mutually exclusive (except, perhaps, under extraordinary conditions, not present, such as very high rates of economic growth).

The preceding table suggests that greater numbers of people are receiving credit and default rates are falling. We are assured that there is no hard evidence that defualt rates are rising. In other words, the market is spontaneously promoting one of the objectives of the Bill.

On the other hand, if costs and risks are increased, economic theory predicts that consumers will experience less access to credit.

2. Reduced over-indebtedness accompanies reduced access to credit

There are a number of ways to assess such impacts. Firstly, a priori (on the face of it), increased risk and cost imposed by the Bill on credit providers, and increased benefits for credit receivers, will result in reduced supply and increased demand, in other words a shortage of credit, which, in turn, means reduced access and a contracting credit market – or, to be more precise, a smaller credit market with less access to credit than there would have been in the absence of the new law.

If we segment the market into three categories of credit receiver: low, middle and high-income, and assume that these categories closely approximate risk-reward categories, we can assess differential impacts on each group. There are formal and sophisticated ways to fragment to market. Informed estimates, based on market realities, suggest that the impact on preferred (high-income) credit receivers will be minimal, and may even be positive thanks to credit being diverted towards them, away from high-risk credit receivers.

Lending to the poor is discouraged by increased risks, such as the prospect of debts being restructured or cancelled, huge penalties and increased compliance costs, which impact disproportionately on the poor. Reduced rate of return will further discourage lending to the poor, if the power to regulate interest rates administratively is used. Finance charge and interest rate caps mean that increased costs cannot be recovered easily.

There will be less propensity to provide finance to the poor, not because they are perceived to be greater credit risks, but because the administration costs of small amounts of credit are the same as for big amounts. It is only at much higher rates of return on small loans that the same amount of return occurs.

Experts we consulted guessed that 10% of middle-income credit receivers and over 30% of low-income credit receivers would no longer be able to access formal credit. One expert argued compellingly that all less-preferred (low-income) credit receivers would be driven from the market because they are all at or near "the margin", where a slightly moved threshold can shift huge numbers of people beneath the threshold if they start just above it.

Since all factors reducing the rate of return to credit providers, such as registration costs and interest rate ceilings, are not in the Bill, probable impacts can only be suggested a an incremental scale. It is estimated that every 1.15% below free market rates results in exclusion from access for 1% of LSM4-7.

What this means in practise is that 40% of the lower end of the market might be driven under the access threshold. Since there is nothing in the Bill that appears likely to increase the supply of credit, or reduce the risk and cost of providing it, it must be predicted that it will result in reduced access.

3. Enhanced remedies

1. Costs

• Fristly, their records turn out to be remarkably accurate,
• Secondly, criticism directed at them for lack of access to credit is misdirected, because they do not grant credit – all they are is a source of informnation for credit-grantors.

1. Credit Insurance

1. Jurisprudential Considerations (Legal aspects)

• sufficient resources are available for implementation,
• the law is realistically enforceable,

• it will be taken seriously by ordinary citizens, especially the lower and middle income, less sophisticated, majority.

• it complies with the letter and spirit of the constitution,
• it satisfies generally recognised principles of good law, such as whether the powers and functions it creates are located in the appropriate organ of state, and
• it complies with the rule of law, natural justice, and due process.

1. Constitutionality

• The Tribunal is a court as defined in the Constitution and is probably unconstitutional because it lacks independence.
• Provisions that discriminate between the national credit register and private registers may be constitutional, but do not necessarily constitute good law – it "does not follow that it is fair or sound public administration".
• The Bill raises "further unanswered questions" (paragraphs 26 to 33).

• clearly constitutional,
• clearly unconstitutional,
• in the grey area,
• probably constitutional, but inconsistent with its values and the principles of good law.

1. Retroactivity

The first and arguably most important section of the Constitution, its "Foundational Provisions", provides that South Africa is "founded" on the "supremacy of the Constitution and the rule of law". One of the central components of the rule of law is that laws must not be retroactive. Apart from the probable unconstitutionality of the provision, retroactivity is jurisprudentially undesirable, and tends to have perverse effects in practice. We did not have the time or resources to consider the practical implications of the provision, and suggest that the Portfolio Committee should not agree to its inclusion in the absence of informed opinion and careful analysis.

It is not clear why the Bill – like others from the DTI, appears to reject the courts (judiciary) as the appropriate locus for civil and criminal dispute resolution of the kind contemplated, and why there is a process of establishing what amounts to a parallel quasi judiciary in the "wrong" branch of government, the executive.

2. The Tribunal

The apparent unconstitutionality of the Tribunal is addressed in the accompanying Senior Counsel Opinion.

 

3. The equality clause

A "grey area" provision is the second issue on which we sought Senior Counsel’s opinion, namely that the Bill provides for fundamentally different rules for the government and private credit registers. This may violate the equality clause (9) and the adminstrative justice clause (33).

The Constitution allows differential rules for distinctive categories of people and institutions, such as distinctions between children and adults, and people disadvantaged by unfair discrimination. It allows differential treatment for government, also for obvious reasons, such as the distinctive status of the army, police and courts, but not where government descends into the same arena as non-government organisations to perform essentially identical functions.

4. Separation of powers

1. Principles of good law

In addition to constitutionality, a comprehensive RIA, as stated, considers the principles of good law generally. Here we consider a few examples.

1. Drafting, structure, style, clarity and accessibility

This is a long and complex Bill. Amongst the principles of good law, especially in modern times and developing countries, are that laws should be in unambiguous, plain and accessible language, that promotes certainty, especially amongst intended beneficiaries and others to whom the law will apply. Paradoxically, the Bill is not a good example of the "plain language" it demands of all documents intended for consumers (§ 64).

Nothing precludes plain language and simplicity in long laws. Indeed, the longer a law, the greater the case for accessibility. One of the best examples of a plain though complex and longer-than-average law is our Constitution.

Every provision in the Bill can be expressed in plain language, and, given its essential objectives – the interests of ordinary consumers, especially less-sophisticated ones, Consideration should be given to having it redrafted. The Law Review Project has drafted many laws for government, and would gladly undertake this task if required.

Many experts are of the view that that good laws have simple, coherent, intuitive and few provisions (see e.g. eminent scholar, Richard Epstein’s, Simple Rules for a Complex World, Harvard University Press, 1995).

2. Monitoring and benchmarks

The essential purpose of a RIA and of good laws generally is to ensure that governments succeed in achieving their objectives. There is a great deal of research to the effect that many or most laws are counterproductive. Indeed, entire schools of economic thought are devoted to studying the phenomenon: "public choice" and "rent-seeking" theory respectively. Legal "positivism" presupposes that laws would have intended effects simply by decreeing what should happen. But the real world is complex and policies turn out to have "unintended consequences". Price control, for instance, would dercree that prices should be lower, but lower prices discouraged supply and encouraged consumption, resulting in shortages. What government have been found to do when price controls are monitored is cause shortages with price ceilings and surpluses with price floors.

If no provision is made for monitoring the effects of a law, and testing those effects against efficacy criteria, governments have no way of knowing – other than anecdotal evidence – whether they are achieveing their goals. One of the questions asked in a comprehensive RIA is whether such perverse effects occur. It is one thing to predict that they won’t or that they will, and another to establish what happens in the real post-regulation world. To that end RIAs establish whether provision has been made for efficacy monitoring.

N provision appears to have been made for monitoring direct and indirect costs and secondary impacts. Specifically, there are no benchmarks according to which success will be regarded as having been achieved. That means the government is unlikely to know as time passes whether it has done the right thing. It will not know, for instance, whether intended beneficiaries are worse off, as it seems they may well be. Harm that can avoided easily may be perpetuated indefinitely.

3. Good law Checklist

There are about 25 criteria recognised as appropriate for laws to be considered jurisprudentially sound. A model Checklist accompanies this document. Every provision in the Bill should be answered in the affirmative. It isn’t, it is either unconstitutional or does not comply with the requirements of good law. We do not include the answer for each provision here. Most are "Yes", but there are too many "Nos" for comfort. Each of these should be corrected or maintained only for exceptional reasons.

2. Way forward

1. Conclusion

• The prohibition of certain kinds of contract is aimed at credit providers, but may impact adversely on credit providers, often more so.
• The informal credit market is being criminalised has implications for many or most South Africans which appear not to have nee considered. The informal market does not entail only small transactions. That the MFRC has only a few of the de facto credit providers on its books may indicate that government does not intend enforcing the law purposefully, or that the Bill will bring credit activity into "the net" for which there is inadequate provision. People driven by the Bill from formal markets may find themselves in the unsavoury underground economy.
• Zero-cost disclosure requirements will lead to a substantial increase in queries and enquiries, by how much being unknown. The cost implications for SMMEs of trying to comply with these provisions appear to be inordinate, and the value of benefits could range form substantial to trivial.
• There is a great deal of knowledge with hindsight regarding the effects of debt counselling and rescheduling provisions and the maintenance of below-market determined interest rates residing with the Land Bank. The general perception is that it was disastrous. It plunged farmers deeply into debt. That experience needs to be considered carefully to ensure that the folly does not repeat itself (see eg Symond Fiske’s definitive retrospective Damage By Debt).
• There is valuable experience with costs to government and compliance costs regarding the Financial Services and Intermediaries Acts and the Labour Relations Act which will give a good idea of what real-world costs might be for this Bill. That information should be considered so that the government can have a better idea of what to expect.

• There are RIAs for corresponding laws in other countries. For the most part they draw attention to costs and benefits, which we mention in this document. More thorough research is desirable though.

1^st Floor Norfolk House P.O. Box 78954

cnr Norwich Close & Fifth Street SANDTON

SANDTON 2146