O'Shea, Paul --- "Consumer Credit Code Disclosure: Does It Work?" [2005] UQLRS 2; (2005) 16(1) Journal of Banking and Finance Law and Practice 5-16

	Home \| Databases \| WorldLII \| Search \| Feedback University of Queensland Law Research Series

This article was originally published in the Journal of Banking and Finance Law and Practice published by Thomson Legal & Regulatory
O'Shea, P, 'Consumer Credit Code Disclosure: Does It Work?' (2005) 16 (1) Journal of Banking and Finance Law and Practice 5-16

Since 1996 all consumer credit transactions in Australia have been regulated by the Consumer Credit Code. The principal means by which the Code purports to protect consumers and prevent market failure is a detailed and prescriptive disclosure regime. There has been little empirical work done on whether or not such disclosure actually improves consumers’ understanding of their credit contracts. By exposing participants to typical consumer credit documents, this research discovered quite poor comprehension of important features of the relevant transactions. Most significantly, there appeared to be little difference in comprehension when the consumers read contracts which complied with the disclosure requirements of the Code, and when they read contracts which did not. These results cast doubt on the effectiveness of the Code disclosure regime.

This article is based on experimental research conducted with funds provided by a University of Queensland Start Up Grant.

“Truth in lending” is more than a catchcry of the consumer movement. It is one of the stated objects of the Consumer Credit Code (“the Code”).[1] Information asymmetries have long been identified as anti-competitive and unfair features of the consumer credit market.[2]

The remedy proffered by the architects of the Code is a detailed and comprehensive disclosure regime. A lack of knowledge and understanding of the likely effects of transactions often leads to their failure to meet consumers’ expectations. This goes to one of the essential elements of contract at common law, namely, the “meeting of minds”. Fix that and, perhaps, there is little more to be done.

Little has been done, however, to test whether disclosure works or, in particular, whether the particular disclosure regime prescribed by the Code, in fact, is effective in meeting the needs of consumers for information about their credit contracts. Some empirical work on the self-professed needs and experiences of consumers has been done by way of surveys but no one in Australia has published research based on experiments to test comprehension of consumer credit contract documents. Nor has any research been published which compares those results against a situation which effectively imitates what would happen if there was no mandated disclosure regime.

The results of this experiment cast doubt on the effectiveness of the disclosure regime mandated by the Code. Further, they raise the possibility that this regime, in its current form, makes almost no difference to consumer comprehension of contract documents. These results lend weight to calls for reform of the current regime and for consideration of other means of consumer protection.

Neo-classical economics assumes that, for markets to provide the most efficient allocation of resources, the parties to transactions have “perfect information” about the relevant products and their cost. When credit consumers have imperfect information they will not be able to make rational choices about products and this will lead to market failure. Trading on consumer ignorance, credit providers will be able to impose terms in consumer credit contracts which favour themselves without cost consequences in the market.[3]

The most primitive compulsory disclosure regime is a legislative requirement that contracts be in writing, the so-called “statute of frauds“ provisions. From the Hire-Purchase Acts to the Credit Acts, each legislative attempt to regulate consumer credit contains such provisions.[4]

Both the Rogerson and Molomby Reports into the consumer credit market observed that consumers had difficulty in understanding the language of consumer credit contract documents, thus effectively undermining the benefits of requiring them to be in writing.[5] The legislative solution adopted was that, both prior to entering into contracts and in the contract document itself, consumers received a statement of prescribed information about the contract. The Credit Acts required it in a schedule and the Code requires it in a financial table which is provided pre-contractually and attached to the contract document.[6]

The Code, at ss 14 and 15, prescribes detailed information to be contained in the financial table as follows (in summary):

Given that the requirements of ss 14 and 15 are substantially similar, credit providers produce the same document and it is common industry practice, particularly in the linked credit situation, for pre-contractual disclosure to occur almost immediately before the provision of the contract document itself.[7]

There are requirements for further disclosures during the “life”, as it were, of the transaction,[8] but this research focuses on the immediate pre-contractual and contractual disclosure.

Work done by Justin Malbon[9] for The Consumer Credit Code Post-Implementation Review Committee has established that consumers use pre-contractual disclosure in different ways depending on the products and that these differences are more statistically significant than those based on different socio-economic demographics.[10]

Malbon’s research method was to use telephone surveys (of almost 1,000 participants) and focus groups. He compared consumer responses to housing loans, personal loans, credit cards and linked credit products. Linked credit produced some of the most interesting results. By linked credit is meant credit which is sold by a trader or supplier of another product so as to facilitate the customer buying that product. The trader or supplier is rarely held to be an agent of the credit provider at common law. The credit provider is therefore often immunised from the consequences of misrepresentations by the trader or supplier despite having had the benefit of new business without the overhead of sales staff and marketing.

Recognising these problems, the Code provides specific redress for consumers of linked credit against both the supplier and linked credit provider, subject to some limitations and defences.[11] The linked credit situation produced:

Malbon concluded that the “truth in lending objectives (of the Code) face a number of hurdles in the linked credit market”.[13] Intuitively, when credit is sold contemporaneously with another product, there is a competition for the attention of the consumer between the features of that product and those of the linked credit. The consumer has not come to the supplier for credit; the consumer has come to buy the supplier’s product, for instance, a car. The consumer can be easily distracted, it is suggested, from the detail of the linked credit product, even if properly disclosed, by the attractions of the supplier’s product which is what, after all, the consumer has come to buy.

These findings about linked credit suggest that it may be a useful place to start investigating the effectiveness of consumer credit disclosure.

The previous research used surveys and focus groups. The former gives less control to the researcher and its author. Even in an anonymous survey, people do not want to admit to having entered into contracts without reading them.

Focus groups give some more control for the researcher, though the responses of the participants become more like “opinions” of what they would like rather than a record of actual behaviour. An interesting feature of the focus groups in the previous research is that it did provide an opportunity to expose participants to “real life” documents.

Field research involving the tracking of actual transactions as they occur, with “exit polls” outside the premises of linked credit suppliers, such as car sales yards, would provide powerful data. This work is however very expensive, demanding substantial resources, a long data gathering period and the co-operation of industry and consumers.

An experiment provides more control for the researcher and can reduce participant control to the bare minimum required for ethical results. This will produce more reliable results. Good experimental design is aimed at overcoming the limitations of the “laboratory” style environment in order to approximate real world conditions. It provides an opportunity for useful research without the costs of going into the field.

The purpose of the experiment was twofold. First, to test whether or not consumer credit contract documents could be quickly and easily understood. Since the documents contain important information about the rights and obligations of consumers, it is suggested that high levels of comprehension of important features of the transaction are required for the disclosure regime to achieve its object of “truth in lending”.

Second, to test whether or not contract documents with disclosure in compliance with the Code were easier to understand than those which embedded the disclosure items in the operative clauses of the contract. If, as discussed above, consumers had difficulty understanding the older style contract documents with important features “buried”, as it were, in the operative clauses of the contract, then extracting such features into the statutory financial table should, if the Code is to achieve its goal, alleviate that difficulty.

These involved obtaining the details of a typical “real life” transaction, obtaining the permission of a leading credit provider to access their standard documentation, re-drafting documents and drafting a questionnaire, obtaining ethical clearance and developing special distracting effects for the experimental environment. The initial task was the drafting of two sets of legal documents. One set of documents consisted of a consumer credit contract with the statutory prescribed disclosure statements as required by the Code (herein labelled “normal” documents). The other set of documents had the disclosure information embedded in the appropriate operative clauses within the consumer credit contract (herein labelled “embedded” documents). The latter is an attempt to proxy for the “non-disclosed” situations which existed in unregulated credit prior to the Code. Both sets of documents could have been used for the same hypothetical transaction because the monetary amounts and descriptions of property and parties were the same.

So as to distinguish between the two types of documents, the normal documents were printed on white paper and the embedded documents on yellow paper. In turn, this would prevent any confusion with the questionnaires when entering the data.

The next task involved setting questions to solicit answers to test the hypotheses. Nine basic questions were asked in the questionnaire relating to information in the legal documents. Some of these reflect the matters required by s 15 of the Code and others relate to the statutory information prescribed in the Code regulations. Not all of them are “key requirements” for the civil penalty regime in s 100 of the Code. The questions are listed below in Table 1. Demographics requested in the questionnaire were age and gender. Except for age, which was designed to be determined on a continuous scale, all other variables were measured in categories.

Table 1 – The Questionnaire

Age:....years

Male/Female

What is the interest rate of the loan?

What is the total cost of the loan?

What is the total amount payable under this loan?

How much will a late payment cost me?

If I pay the contract out early, will I pay extra fees?

Did I get consumer credit insurance with this loan?

Can the credit provider’s employees come onto my private property to repossess the car without my consent?

Can the credit provider add other credit fees and charges after I sign this contract?

Was there any commission paid on this loan?

Once the documents and questionnaire were designed ethical approval for the study was obtained. The University of Queensland Behavioural and Social Sciences Ethical Review Committee granted approval. Given that participants were unable to be identified from their questionnaires, anonymity was guaranteed as was confidentiality and privacy.

Various distractions, simulating those of a normal car yard environment, were created whilst some participants filled out their questionnaires. The two types of distractions were audio and visual. The audio distraction was a continuously looping tape recording of up beat dance music underneath a male voice talking constantly and rapidly about the attractive features of different motor cars. The visual distraction was a slide show of pictures of car yards and impressive looking vehicles projected onto a screen in front of the participants.

A small pilot study was carried out to pre-test the questionnaire for clarity of wording and clear instructions. Introductory psychology students were recruited for the pilot study. No other sampling criteria were imposed. Recruitment was achieved by placing an advertisement on a website asking for participation in the study, which included a payment of $10.00 at completion of the questionnaire. Twenty-four students signed up to be part of the convenience sample and only 16 turned up to be part of the research.

All participants received a “set” of documents typical to the purchase of a used car with linked credit. These included:

The participants were not told which of these documents were the most important. They were given 15 minutes to read the documents and then a further five minutes with the documents still in hand to answer the nine questions. Audio and video distractions were run during the pilot, though these were improved for the main study experiments with more pictures, more prominently displayed and louder music.

Feedback was sought from the pilot group about the experimental design, in particular, the time needed to answer the questions, and no participant indicated that they needed more time.

Question 2 “What is the total cost of the loan?” seemed troublesome because only one student answered this correctly. Therefore, the wording was assessed again for the main study. The reworded question was, “What is the loan costing you?”

The sampling method for this research was influenced by convenience and achieved a substantially cross-sectional result. The sample consisted of students from the University of Queensland. In 2003, the combination of minimum entry requirements and demand for places for undergraduate courses at the University of Queensland resulted in students from only the top 27.7% of the Year 12 graduating class being allowed to enrol.[14] The age of participants, ranging from 16 to 37 with an average of about 20, is also not comprehensively representative of the general population but is quite typical of consumers buying used cars on credit and, particularly, their first car.

Recruitment was achieved in several ways. Some announcements were made in classes, the website advertisement continued to run and, immediately prior to each experimental session, a research assistant approached students at random on the campus grounds. They were asked to participate in an experiment and be paid $10 for such participation. The research assistant gave these students a brief outline of the experiment and instructions about times and venues.

When students arrived at the venue, they were asked if they had participated in the experiment before and if not they were given an information sheet to read and asked to sign the consent form. No names were recorded twice. The sample consisted of 283 students. However, one batch of contracts was missing vital information, therefore 68 questionnaires were excluded from the data analysis. Hence n = 215.

Nonetheless, even though those questionnaires could not be included in the data analysis, they did provide some valuable information. In the pilot study for this research, it was noted that question 2 “What is the total cost of the loan?” was poorly answered, that is only 1 student out of 16 (6.3%) gave the correct answer. The question was changed to “What is the loan costing you?” However, of the 68 questionnaires missing vital information only 3 participants (4%) answered “What is the loan costing you?” correctly. Consequently, because there was no difference in understanding of the two questions, the remaining 215 participants answered the question “What is the total cost of the loan?”

The experiment was conducted in batches or sittings of about 30 participants at a time in conditions similar to those of the pilot study. Data was collected as answers to questions relating to the normal disclosure and embedded contract details. The collection of data was undertaken under two conditions: The first was a noise-free or exam-like environment; the second with the distractions designed to proxy for the used-car yard environment. Prior to data collection, the audio and visual distractions were set-up in various classrooms at the University of Queensland’s campuses at St Lucia and Ipswich.

This data analysis is divided into three main sections. The first relates to results for the whole sample. It includes frequencies, measures of central tendency and a t-test to ascertain whether those answers of the participants who read the disclosure contracts were significantly different from those answers of the participants who read the embedded contracts. Further, gender and age are evaluated to check for an association with comprehension. In the second and third sections, the analysis focuses on the different types of contracts, that is, the new disclosure contract and the embedded contract respectively. The descriptive statistics used in this section include frequencies, measures of central tendency, and t-tests. T-tests were used to determine whether or not conditions, which were distraction and distraction-free, affected participants’ level of comprehension.

Table 2 – Correct Answers to Individual Questions for the Whole Sample
Questions	Frequency	Percentage
Q1. Interest rate on the loan.	209 students	97.2%
Q2. Total cost of loan.	16 students	7.4%
Q3. Total amount payable under loan.	166 students	77.2%
Q4. Cost of late payment.	142 students	66.0%
Q5. Extra fees for paying contract out early.	148 students	68.8%
Q6. Consumer credit insurance.	110 students	51.2%
Q7. Repossess car from private property.	59 students	27.4%
Q8. Other fees and charges.	130 students	60.5%
Q9. Was there any commission paid?	147 students	68.4%

= 2.9% of the difference in comprehension can be explained by the differences in the documents.

As previously mentioned, there were 112 participants who read the disclosure contracts.

6. The total number of correct answers for each of the 9 questions is listed in the table below:

Table 3 – Correct Answers to Individual Questions for Disclosure Contracts
Questions	Frequency	Percentage
Q1. Interest rate on the loan.	110 students	98.2%
Q2. Total cost of loan.	5 students	4.5%
Q3. Total amount payable under loan.	92 students	82.1%
Q4. Cost of late payment.	71 students	63.4%
Q5. Extra fees for paying contract out early.	80 students	71.4%
Q6. Consumer credit insurance.	63 students	56.3%
Q7. Repossess car from private property.	36 students	32.1%
Q8. Other fees and charges.	69 students	61.6%
Q9. Was there any commission paid?	85 students	75.9%

7. Of the 112 participants who read the disclosure contracts, 62 read their contracts and completed their questionnaires under a no noise condition (mean = 5.3; sd = 1.4), whereas 50 read their contracts and completed their questionnaires with audio and visual distractions (mean 5.70; sd = 1.3). In order to check for any difference in levels of understanding between participants who read their contracts with distractions to those who read their contracts without distractions an independent t-test was carried out. The variances in both groups were equal (p = .54). The t-test was not significant (t = -1.47; [110 df]; p = .15). Therefore, there was no difference in the level of comprehension between participants who read the disclosure contract in an environment with distractions from those who read the disclosure contract without distractions.

In the sample where students read the embedded contracts, there were 103 participants.

5. The total number of correct answers for each of the 9 questions is listed in the table below:

Table 4 – Correct Answers to Individual Questions for Embedded Contracts
Questions	Frequency	Percentage
Q1. Interest rate on the loan.	99 students	96.1%
Q2. Total cost of loan.	11 students	10.7%
Q3. Total amount payable under loan.	74 students	71.8%
Q4. Cost of late payment.	71 students	68.9%
Q5. Extra fees for paying contract out early.	68 students	66.0%
Q6. Consumer credit insurance.	47 students	45.6%
Q7. Repossess car from private property.	23 students	22.3%
Q8. Other fees and charges.	61 students	59.2%
Q9. Was there any commission paid?	62 students	60.2%

There was little support for the two hypotheses in these results. On the first hypothesis, whilst overall comprehension of some questions (eg interest rate) was quite high at 97.2%, most questions were answered correctly by only about 2/3 of participants with the overall mean being 5.3 out of 9. What is most disturbing is that, arguably, the most important question, about the total cost of the loan, was answered correctly by only 7.4% of participants.

Not only did the financial table present comprehension problems for most participants, but so did the statutorily prescribed statement of rights and obligations. Only 27.4% of participants understood that the credit provider could not repossess mortgaged goods from private property without consent or a court order. This issue is specifically addressed in the prescribed statement supplied to participants using the Code-compliant documents.

These mixed results for overall comprehension are interesting but not, in our view, as important as the result for the second hypothesis. This compared comprehension of Code compliant documents with embedded contract documents as proxies for pre-Code documentation. Whilst there was a barely statistically significant improvement in the mean for Code-compliant disclosed contracts over the embedded contracts (5.51 as opposed to 5.02), only 2.9% of this variance is explained by the difference in the documents. The rest is, therefore, attributable to other factors.

In other words, the Code disclosure regime makes very little difference in the comprehension levels of important features of the transaction for the participants in this experiment.

The first and most obvious limitation is that this is an experiment. The participants lack the “real world” incentive to study the documents because they are not actually going to affect their lives the way they would if they were buying a used car on credit.

Taking this view is, of course, to dismiss the willingness of all experimental participants to co-operate, to “earn“ their $10 fee and to contribute to the expansion of knowledge which hopefully comes from experimental research. It further tends to dismiss all social science research which is based on experiments rather than surveys or field or action research. Luckily, it is often the case that “phenomena discovered using hypothetical choices have been reproduced in studies using actual choices.”[15]

There has been theoretical work done which questions the incentives for “real world” consumers to either absorb or act on information supplied to them about their transactions. Michael Trebilcock argues that if the value consumers put on information in terms of “making a better choice about what goods or services to buy and on what terms” is not high then “...in a world of information overload, consumer protection instruments that actually generate information that is costly for consumers to interpret or access may be counterproductive”. He suggests that this may be particularly so for statutes “that mandate detailed disclosure of contents or ingredients, complex details of the price, terms and conditions of a transaction or very specific caveats about the use of the product”.[16]

By highlighting the differences between the “impulsive self” and the “planner self” in the human psyche, behavioural economists also suggest that consumers will, either consciously or unconsciously, sometimes simply ignore information disclosure and act in ways which suit their short term prejudices.[17]

The second limitation is that the results also indicate no statistically significant differences in comprehension levels can be attributed to the presence or absence of the artificially created distractions. This gives rise to two possible explanations and consequent courses of action. Either the distractions were insufficiently realistic to imitate the reality of the used-car linked credit situation or, alternatively, such distractions, in the real world, make little difference to comprehension levels.

The third limitation is that the sample is clearly not truly representative of the education levels and potential of the general population. The “elite bias” however, if anything, tends to strengthen rather than weaken the results.

The second limitation on this experiment could be addressed by intensifying the distractions. Further, there is likely to be a relationship between time spent with the documents and comprehension levels. Thus, other types of loan transactions, such as home loans (which the respondents to Malbon’s survey claim they spent more time reading) could produce better comprehension results if the participants are given more time.

Many other transaction types such as investment or insurance, which are the subject of legislatively mandated disclosure regimes, could be the subject of similar experiments in order to test their effectiveness.

The limitations of the sample population could be overcome by more resource-intensive expansion of the sample to remove the elite bias and more closely to reflect the age and educational diversity of the general population. More expansively, and in response to the first limitation, the field or action research option discussed above could be pursued so as to self-correct for any perceived experimental sample bias or lack of participant motivation. Tracking and testing large numbers of real world consumers making actual choices in the market place is, of course, the most empirically sound approach.

The technology is available to examine more closely, again in the experimental environment, the cognitive dynamics of document comprehension. Using a combination of miniature video cameras attached to headsets, digital video recordings and specialist software, researchers can track where participants are looking as they seek the answers to particular questions. We can thus achieve a better understanding of exactly where consumers think the answers should be.

This last line of research will be useful in improving the existing disclosure regime. Elizabeth Lanyon says that “The narrow original aim of ‘truth in lending’ which focussed on pricing choice between products is itself now being re-evaluated both as to whether disclosure aims should be broadened and also what level of detail or precision is most helpful to consumers.” She also reports that the Uniform Consumer Credit Code Management Committee is working on adopting the “Schumer box” format for simplified disclosure in consumer credit contracts.[18]

Iain Ramsay says that we should continue the “process of adaptation and experimentation” with disclosure and that:

[I]t would be dangerous...to use information overload as a general rationale for reducing information without analysis of when the disclosed information will be used by consumers. While individuals may absorb a limited amount of information at the point of purchase they may be very attentive to information on contract terms, their rights and sources of advice if trouble arises in relation to the credit contract.[19]

Lanyon suggests that “a thorough re-evaluation of the centrality of disclosure needs to occur” and she argues for new strategies including improved consumer redress mechanisms.[20] One such mechanism may well be unfair contract terms legislation along the lines of that recently adopted in Victoria and the United Kingdom.[21] Although the former does not apply to credit, its adoption by other States and extension to credit is being championed by consumer agencies and considered by the Standing Committee of Officials of Consumer Affairs.[22]

Linking the potential failures of disclosure to the need for substantive regulation of contractual terms is not a new idea.[23] It is however an idea which lost its influence in the 1990s as demonstrated by the reliance on disclosure as the principal means of consumer protection in the Consumer Credit Code. The results of this experiment indicate that it is an idea whose time has come again.

[∗] BA(Hons), LLB (UQ), MSc (UO) GDLP (QUT), Lecturer, TC Beirne School of Law, University of Queensland, Solicitor of the Supreme Court of Queensland and the High Court of Australia.

[∗∗] B Nurs (Hons) (QUT) PhD (UQ), Key Centre for Human Factors and Applied Cognitive Psychology, University of Queensland.

[2] Rogerson, Report of the Standing Committee of Commonwealth and States Attorneys-General on the Law Relating to Consumer Credit and Money Lending (Rogerson Report) ( South Australian Government Printer, Adelaide, 1969) p 22-25; Molomby, Report to the Attorney-General for the State of Victoria on Fair Consumer Credit Laws (Molomby Report) (Victorian Government Printer, Melbourne, 1972) p 12.

[3] Duggan A, “Economic Analysis of Standard Form Contracts: An Exposition and a Critique” in Cranston, R and Schick A (eds), Law and Economics, (Department of Law, Research School of Social Sciences, ANU, Canberra, 1982) p 146.

[7] Duggan A and Lanyon E, Consumer Credit Law (Butterworths, Melbourne, 1999) at [4.3.3].

[9] Malbon J, Taking Credit, Report for the Consumer Credit Code Post-Implementation Review, (Tasmania, Department of Justice and Industrial Relations, September 1999), http://www.creditcode.gov.au viewed 4 February 2005.

[12] Malbon J, “Shopping for Credit: Empirical Study of Consumer Decision-making” (2001) 29(1) ABLR 44.

[15] Thaler R and Lowenstein G, “Intertemporal Choice” in Thaler R, The Winner’s Curse: Paradoxes and Anomalies of Economic Life (The Free Press, New York, 1992) p 95.

[16] Trebilcock M, “Rethinking Consumer Protection Policy” in Rickett C and Telfer T, (eds) International Perspectives on Consumer’s Access to Justice (Cambridge University Press, 2003) pp 70 –75.

[18] Lanyon E, “Changing Direction? A Perspective on Consumer Credit Regulation”, Keynote Address, Australian Credit at the Crossroads Conference, 8 November 2004, Melbourne, p 16, http://www.consumer.vic.gov.au viewed 4 February 2005.

[19] Ramsay I, “Consumer Credit Regulation as ‘the Third Way’”, Keynote Address, Australian Credit at the Crossroads Conference, 8 November 2004, Melbourne, p 15-16, http://www.consumer.vic.gov.au viewed 4 February 2005.

[21] See s 32 et.al of the Fair Trading Act 1999 (Vic) as amended by the Fair Trading (Amendment) Act 2003 (Vic) and the Unfair Terms in Consumer Contracts Regulations 1999 (UK).

[22] Standing Committee of Officials of Consumer Affairs, Unfair Contract Terms: A Discussion Paper (2004) p 58; for a submission which argues for application of unfair contract terms legislation to consumer credit see Howell N and O’Shea P, Submission to the Standing Committee of Officials of Consumer Affairs Unfair Contract Terms Working Party (Centre for Credit and Consumer Law (Qld)), p 13, www.fairtrading.qld.gov.au/oft viewed 28 January 2005.

University of Queensland Law Research Series

O'Shea, Paul --- "Consumer Credit Code Disclosure: Does It Work?" [2005] UQLRS 2; (2005) 16(1) Journal of Banking and Finance Law and Practice 5-16