ISSN: 1204-5357

All submissions of the EM system will be redirected to Online Manuscript Submission System. Authors are requested to submit articles directly to Online Manuscript Submission System of respective journal.

The missing credit information system in Hungary

  1. Deputy Governor, Magyar Nemzeti Bank (the Hungarian National Bank), Hungary Postal Address: Szabadság tér 8-9, Budapest, Hungary, H-1054 Author's Organizational Website: Email: [email protected]
  2. Lecturer, Finance Faculty, International Business School, Budapest, Hungary Postal Address: Tárogató út 2-4, Budapest, Hungary, H-1021 Author's Organizational Website:
    Email: [email protected]
Related article at Pubmed, Scholar Google

Visit for more related articles at Journal of Internet Banking and Commerce


Hungary is one of a few countries that do not operate either a national, non-profit, or a profit oriented business-based complete mandatory credit information system (i.e., credit register or credit bureau). In its absence, the Hungarian banks have not enough information on their borrowers to evaluate their creditworthiness, to apply behavioural scorings or proper risk management and credit pricing tools. The debate on the necessity of a complete credit information system has almost a decade long history in Hungary. Until now, despite all the professional arguments, the counter arguments from the point of view of personal data protection proved to be stronger. The lack of the complete credit information system became even more painful during the recent financial crisis, when the quality of the retail credit portfolio significantly deteriorated. This paper presents the most important milestones of the institutional history of the struggle for the complete credit information system and analyses the tendencies on the Hungarian retail lending market, with special regard to the lack of it



banking; credit information system; retail lending; Hungary


Hungary was hit particularly hard by the financial turmoil of 2007–2010. The banking sector and the FX money market have experienced a serious liquidity crisis, the government bond market collapsed, GDP declined by 7% and the Government was forced to seek extraordinary financial support from the EU and IMF to the amount of EUR 20 billion. Nevertheless, the country weathered the crisis well, with the banking sector managing to avoid solvency problems and none of the banks had to be bailed out. In 2010, a slow recovery has started, assisted by disciplined fiscal and monetary policies. With the crisis largely behind us, we can now summarise the lessons and the main elements of the misguided economic policy that led to it.
Those elements included, among others, a sharp expansion of foreign currencydenominated credit to unhedged households in 2004–2008. Banks extended high LTV loans to low-income borrowers. The process had a special trait: all the loans were extended to the households without any credit history. This was simply the consequence of the fact that a complete mandatory credit information system for the households does not exist in Hungary, and that is why behavioural scoring, or credit history analysis, has been excluded from the banks arsenal.
In this paper, we first summarise a long history of the non-existing credit information system and then analyse its consequences for the retail loan market.


Hungary is one of a few countries that do not operate either a national, non-profit, or a profit oriented business-based complete mandatory credit information system (i.e., credit register or credit bureau). A partial credit register has been in place in Hungary for more than fifteen years by now – BISZ Ltd (owned by Hungarian financial institutions) was founded in 1994. This partial database covers the most important data about all corporate loans (both performing and non-performing), but only the non-performing personal loans. The history of the still non-existing complete personal credit information system spans almost a decade. In this chapter, we summarise the most important milestones of this long and seemingly never ending story.
The two-tier banking system was established in 1987, and the newly created banks had very poor credit risk management. Bank lending was similar to the ‘blind flying’, because non-existing credit information made it impossible to evaluate fairly future bank’s customers. Due to ‘bad luck’ (the economy endured a deep recession at the beginning of the 1990s), ‘bad regulation’ (even sound companies were forced to the bankruptcy as a result of the introduction of the bankruptcy law) and ‘bad banking’ (accumulating risky portfolio), the collapse of the banking system was inevitable: in 1993 the Hungarian ROE was minus (!) 100%.
The total capital loss and the subsequent recapitalisation brought about the requirement for banks to create a database that could be used to reduce credit risk. The creation of such a database followed the development of the necessary legal framework for its mandatory use, as well as the permissions for, and the scope of, managing its data. On 15 February 1994, eight banks established BISZ Ltd1 (Interbank IT Services), a private corporation, for the development of the credit information database. BISZ was expected to design and operate the Interbank Debtor and Credit Information System (IDCIS)2.
The corporate sub-system of the IDCIS was launched on 28 June 1995, containing full records on corporate and SME loan contracts. By mid-1996, all of the banks, savings cooperatives and credit unions operating at the time had subscribed to the service. Data submission to the system was continuous, and by the end of 1996 the records had become virtually complete. Now it covers the data files of 180 000 Hungarian corporations with almost 1,4 million contracts.
From 1998 the new banking law forced BISZ Ltd to launch a new “retail” sub-system, now for non-performing personal loans. This sub-system contains all natural persons who have applied for a loan but ‘have not fulfilled their payment obligation for a period of more than 90 days, exceeding the amount of minimal wage’. Borrowers registered in the database were excluded from the market for 5 years.
Since the system is compulsory, banks can use it as a primary filter: if somebody were registered in the ‘bad loan book’ his/her credit application would be rejected. However, this partial retail sub-system is incapable of functioning as a database for behavioural scoring, similar to credit bureaus operating in the Netherlands, Germany, Czech Republic or the Fair Isaac system in the US. As a consequence, Hungarian banks are still ‘flying blind’ when they lend to the households, as will be analysed in more details at the second part of this paper. The ongoing debate in Hungary is about the completeness of the retail database: should it be a ‘negative list’, i.e. listing only defaulters, or should it be a ‘positive list’, i.e. listing all the debtors with their full credit history.
Discussions of a similar type taking place in the world have a somewhat different ‘privacy’ focus. The basic question is whether a ‘voluntary’ or a ‘compulsory central’ credit information system should be established. In the discussions, usually no distinction is made between ‘private debtors’ and ‘corporate debtors’. Academic publications generally analyse the effect of ‘public’ information available on ‘the’ debtors.
Mandatory credit registers are operated in many countries, and the type of debtors they cover can only be established after their detailed survey. At the time of establishing BAR nobody argued that mandatory information sharing has a goal of ‘improving social welfare’; however, how banks provide (in what quality) and use this information is a different issue.
The essence of the Hungarian dispute is about whether or not a ‘positive list’, i.e. a complete credit information system violates citizens’ rights related to the protection of personal data. The right of personal privacy means that everyone might control their own data themselves. So one has the right not only to request keeping one’s personal data confidential to the highest possible degree, but also to actively determine the handling of one’s personal data and to control the content and distribution of their personal data kept by the state or the market players (Court of Constitution Resolution No. 15/1991). Hence, a nation-wide debate concerning the credit information system was about the introduction of the positive list and of the protection rights for personal data.
In 2002 the Hungarian Banking Association has proposed for the first time to establish a ‘complete credit information system’, i.e. a ‘positive list’ of debtors. Shortly afterwards, the Magyar Nemzeti Bank (MNB, the Central Bank of Hungary) published the first professional analysis on the topic (see Árvay, Dávid and Vincze, 2002). Its authors made the usual (though not uniformly accepted) distinction between the credit registers and credit bureaus. In their interpretation, a credit register is a ‘public’ organization, regulated and operated by some authority on a non-profit basis, whereas a credit bureau is usually a ‘private’ profit oriented business.
The study not only compared international credit registers and credit bureaus, but also provided a detailed survey of individual countries,. The authors justified the necessity of a similar Hungarian database to develop banks’ credit risk management, clarifying why such organizations are of the vital importance for the Hungarian banking sector in its support of the growing retail lending business.
This early initiative was rejected by both the financial supervisor (in Hungary the Financial Supervisory Authority, HFSA, is as independent as are its counterparts in the UK or Sweden) and the Hungarian Parliamentary Commissioner for Data Protection and Freedom of Information (from here, the Ombudsman). At a conference held at the International Training Center for Bankers in January 2003, the Ombudsman’s Deputy Commissioner clearly articulated his organization’s opinion that the right to the protection of personal data is a fundamental right under Article 59 of the Constitution. A fundamental constitutional right might only be restricted to enforce other basic rights, if it means public interest that:
(i) makes justified a limitation of the fundamental constitutional right;
(ii) the specified goal cannot be achieved in any other way; or
(iii) the restriction is suitable for reaching that goal.
In the Ombudsman’s argument, improvement in creditworthiness is not a constitutional goal that could override the right to the information’s privacy. Economic necessity is not a sufficient argument. As well, the cases of misuse of large registers data banks (that is, their use for other purposes than that when the original database was created) are relatively common.
The above arguments have not changed with the passage of time; the opponents of the positive debtor list have always emphasised that such a list not only contradicts the provisions of the Data Protection Act but also raises substantial constitutional issues. According to their argument if personal data are provided mandatory for the lender it would restrict (limit) privacy rights.
The Ombudsman had on several occasions referred to the fact that complete retail credit information systems do not exist in all countries, but that many states have a voluntary debtor list, which is not in the conflict with the constitutional privacy provisions. Regretfully, a thesis that has been proven by several theoretical and empirical studies, namely that voluntary institutions do not have an unequivocal effect on either consumer protection or the increase in market competition, have not influenced the Ombudsman’s opinion. Consequently, such critical arguments have to be extended in support of a positive debtor list, which makes intervention via legislation essential. It is critical to resolve whether public interest in this case justifies a limitation of the fundamental privacy right Of course, it is also necessary to prove that the tool selected for the achievement of this objective – in the present case the positive list – is suitable for reaching the desired goal.
After a temporary quiet phase, the discussion has sparked up again in 2006, precipitated by a slowdown in the retail market’s growth, a frightening degree of private indebtedness, perceptible weakness of the retail leasing market, a steady increase in the volume of non-performing loans and the radical improvements in debtor rating and debtor selection methods. All these critical events have inevitably brought regulators (the HFSA, the MNB and the Ministry of Justice) and the industry together (even a business line executive of the largest Hungarian retail bank gave a whole page interview to a leading economic newspaper) – a positive debtor list, that is a complete credit information system is a necessity after all!
At the beginning of 2006, a path-breaking new study was published on the HFSA’s website, expressing a somewhat softened view by this institution. Dialogues started again, holding out hopes of rapid progress for the interested parties. Amusing intermezzos made the debate vivid: the Parliamentary Commissioner for Human Rights suggested a ‘positive list resolution’ in January, while at the same month the Data Protection Ombudsman denied the above-listed Parliamentary officer’s right even to interfere in the discussion.
Regretfully, the 2006 debate have not yielded any new result. Meanwhile, one of the most active credit information providers, TransUnion-CRIF3, which already had some subsidiaries in the Central and Eastern European countries entered the Hungarian market and tried to build a voluntary positive-list based database, but the legislation did not give way to this initiative. Banks, with the support from the Banking Association, the MNB and the HFSA, tried to form a lobby to achieve a path-breaking modification to the legislation. However, the two largest players with data-rich in-house databases (one, a large retail bank OTP Bank, is a Hungarian ‘national champion’; the other is a local subsidiary of Cetelem, France) did their best to place as many obstacles as possible on the way, as they did not want to lose their competitive advantage.
The supporters of the complete credit information system emphasised that these kinds of systems are widely used in the developed countries, and even in some emerging economies. Complete credit registers exist, for example, in Austria, Belgium, the United States, the United Kingdom, the Netherlands, Ireland, Norway, Italy and Sweden. (In reaction to this argument the Ombudsman said that a kingdom as an institution also exists in many of these countries but it is missing in Hungary…)
A common argument in the dispute is that a positive debtor’s list allows for more accurate assessment of risks for the banks. By itself this argument is not particularly compelling; as it is not in the public interest to restrict a fundamental human right. Another reason cited in many working papers is that a more accurate risk assessment could result in the enhance security of the banking sector, increase in the financial stability, thus improving public welfare. The decrease in the number of defaults would also contribute to the consumer protection.
Several recent papers stress the importance of putting a constraint on further indebtedness. And, last but not least, many defenders of a positive list argue that prices, i.e. interest rates, might decrease significantly as a consequence of its introduction. This argument is somewhat closer to the ‘public interest’ but is too indirect to limit the fundamental constitutional right, and it currently cannot be supported by the empirical evidence. During the latest financial crisis, when the ratio of NPLs within Hungarian retail portfolios exceeded 10%–15%, this argument was presented again, receiving a somewhat wider support.
At the end of 2006, the Prime Minister of Hungary set up a professional committee entrusted with investigating main problems of the retail banking services. Within just three months the so called Várhegyi Committee4 has prepared a detailed analysis of the crucial weaknesses of country’s markets and institutions. An important part of the report was a proposal to found a complete information system by enlisting the necessary regulatory steps. The report recited all the ‘pros’ and ‘cons’ and concluded that a complete credit information system would increase social welfare.
The report was widely discussed with several working groups being set up, and the Ministry of Justice started draft a new law. At the beginning of 2007, an international conference on credit information systems was organised at the Hungarian International Training Center for Bankers. The organisers invited professor Jappelli, one of the foremost academic scholars of this subject, who reported on the status of empirical research performed in several European countries in the past 15 years. As well, several regulators and supervisors from the European countries were invited to describe their experiences. While representatives of the MNB and HFSA did support the initiative of establishing a new credit information system, the Ombudsman did not accept the invitation of the organisers, stating that all the ‘cons’ have already been repeated by him many times before. In the early summer of the same year the HFSA organised a similar conference, where a possible framework for a voluntary system had also been discussed.
Despite the unanimous voice of the industry, no breakthrough has been reached, as the new law for the establishment of complete credit information systems was not further advanced. It was the financial crisis that resulted in the new substantial step. A financial crisis hit Hungary in the post-Lehman period in October 2008, and the country vulnerable to shocks had to apply for an IMF loan. In the very first Letter of Intent the necessity of the ‘positive debtor list’ was articulated: ‘Over the period covered by the IMF program, financial sector regulation and supervision will be further strengthened. Measures include:
(i) introduction of a positive credit registry for households;
(ii) modification of the Central Bank Act to allow the MNB to request individual but unidentifiable data to adequately analyse credit risk;
(iii) enhanced regulation of insurance and credit brokers and their products;
(iv) introduction of maximum Loan-To-Value Ratio requirements for new mortgage loans; and
(v) close monitoring of banks’ foreign exchange exposures.’5
The corresponding legislation was drafted in just one month (as by that time numerous previous drafts had already been piled up in the drawers of the Ministry of Finance), and after a month of intra-governmental coordination it was put forward to the Cabinet. At the end of October, the Government accepted the draft, and in December 2009 it put it on the Parliament’s Agenda.
The draft legislation proposed the establishment of a mandatory credit register. The register would contain new and already existing retail loan contracts; i.e., the problem of the ‘original upload’ would thus have been solved. The operation of the register would have met the most strict data protection requirements under the supervision of the HFSA. The register could have been operated by any professional profit oriented organisation that complies with several criteria – the most restrictive among was that the register should cover 70% of all country’s financial institutions and 80% of the total household debt. (As a consequence, probably one credit register would have been set up – and it really would have hurt competitiveness.)
However, the Parliament postponed the discussion of the draft legislation until 2010 and, after the spring 2010 election, the new Government removed it from its agenda. Moreover, the only regulatory change that the new government made has in fact weakened the information content of the compulsory credit information system. In December 2010 the Parliament passed a law that have decreased from five to just one year the system’s memory, i.e. the exclusion from the system has to be done in a year in case of loan recovery.
Nevertheless, there is at least a partial happy ending to this story. While the legislation reached a dead end, the market’s effort to set up at least a voluntary positive debtor list has been fruitful. In 2009 two business organisations6 have started their credit reference services with the HFSA’s permission. These organisations are allowed to collect information on a voluntary basis, and, of course, only about the new loans, i.e. it will take some time to collect a substantial volume of relevant data. Both organisations advertise themselves as ‘the first consumer credit information sharing service’ in Hungary, although none of them could actually yet provide a sufficient information – they are both in the process of data gathering and attracting customers. The services aim to enable both consumers and banks to benefit from the advantages of positive credit histories.


According to Jappelli and Pagano (2006), the advantages of information sharing among credit market participants are as follows:
(i) it improves a bank’s knowledge about borrowers’ characteristics, which in turn reduces adverse selection and facilitates more precise credit pricing;
(ii) it reduces banks’ informational rents (i.e. reduces the possibility of realizing extra profits on the exclusively available information, that is by definition increase the competition);
(iii) it imposes a discipline on the borrowers; and
(iv) it eliminates the incentives to over-borrow with the multiple lenders.
In an earlier article (Király and Mérő, 2008), we have analysed the effects of the lack of a country’s credit information system on the Hungarian retail banking market up till 2007. We have found that the absence of all the four above-mentioned information sharing advantages could have been observed. At that time, we had information about the US sub-prime crisis, but as we were still in the pre-Lehman period, a massive international financial and economic crisis seemed unlikely. As the Hungarian financial markets were fairly underdeveloped, the banks had no toxic assets in their portfolios. At that time, the generally held view was that emerging markets could somehow ‘decouple’, remaining relatively unaffected by the US crisis.
Since then the global crisis has escalated dramatically and Hungary became one of the countries severely hit by it. Now, it is worthwhile to re-examine the tendencies in the Hungarian retail lending market, especially in regard to the continuing absence of a proper credit information system.
With the exception of the Hungarian ‘national champion’ OTP Bank, the monopolistic retail banking market player until 1990, all the dominant retail market participants entered the Hungarian market in the 1990s or early 2000s. As they started to develop their retail lending business gradually, the retail lending-to-GDP ratio was below 5 per cent even in 2000. Against this background, it is not surprising that despite the sharp increase in the retail lending volumes in the Hungarian banking market before the financial crisis, the retail lending-to-GDP ratio remained much lower than that of the euro-zone countries or of the US banking system (see Chart 1).
These characteristics of the Hungarian retail lending resulted not only in a high level of concentration and a low level of competition, but also in the lack of a credit history in general and, in particular, a decline in the debtors’ propensity to repay their loans. Banks cannot have proper information on the creditworthiness of their clients because:
(i) they have a short history and relatively few clients in the Hungarian retail market; and
(ii) the majority of clients are new debtors with no or with a very limited repayment history.
In this environment, the availability of credit information is especially important, as – save for OTP –no Hungarian bank has a long-term broad information base for the retail lending.
Despite the low retail credit-to-GDP ratio, due to higher interest rates, typically shorter repayment periods and relatively low income of the households, the debt service to disposable income ratio of Hungarian households had reached that of the overall eurozone by the mid 2000’s and even slightly exceeded it before the crisis hit Hungary. An extraordinary high growth rate of both housing and consumer credit was coupled with the serious indebtedness of the Hungarian households and a lack of financial literacy. Prior to the crisis, households have had little experience in taking credit risk, knowing very little about the factors potentially influencing their repayment burden. At the same time, the low credit-to-GDP ratio and the high demand for loans have caused the illusion that there was a huge potential for further credit growth. The lack of information on clients’ repayment capacity contributed to this illusion and opened a way to the irresponsible lending practices. Banks wanted to sell their credit products and preferred not to disclose all the related risks with the sufficient details.
Another characteristic of the retail lending in the pre-crisis period was that, due to the large interest rate differential between forint-denominated loans and foreign currency loans, there was a permanent shift in lending towards foreign currencies. Chart 2 compares the interest rate differential between the euro-denominated retail lending in the euro-zone with the forint-denominated retail lending in Hungary .
Chart 2 Interest rate differentials between the forint-denominated retail lending in Hungary and the euro-denominated retail lending in the euro-zone; expressed as the annual percentage rates of charge (APRC)
This shift towards the foreign currency lending did not increase banks’ foreign exchange risk, as they funded the loans in foreign currencies or hedged themselves in the swap market; however, it resulted in a significant credit risk. Banks were exposed to the risk that, if the forint were depreciated, the increase in households’ repayment burden could have lead to a rise in the ratio of non-performing loans, as actually happened during the crisis.
However, during the pre-crisis period there was a sustained forint appreciation, which has led to the illusion that the Hungarian forint is stable and not subjected to the depreciation. This illusion has led some households to maximise their monthly instalments that they could repay, not taking into account the potential increase in repayment burdens due to the depreciation of the forint. This shift was so strong that forint lending stopped almost completely (see Chart 3).
Fuelled by the ample lending, Hungarian households’ consumption and house purchase became higher than the population could afford. The price of it was that the borrowers have taken high foreign exchange risks. If the forint were ever significantly depreciating, the borrowers could massively default on their loans.
Because most of the loans were covered by the mortgages, the banks have thought that they are in a safe position. Low delinquency rates also contributed to this illusion. (In 2006, housing loans accounted for 59 per cent of retail loans, with consumption loans covered by mortgage accounting for another 18 per cent.)
Consequently, by 2006, it has became obvious that the Hungarian households were unaware of the risk they have taken, that the banks did not properly inform them , and that in the event of an economic downturn the Hungarian household indebtedness could cause very serious problems indeed.
The lack of a proper credit information system played a role in the acceleration of the indebtedness process. As banks’ knowledge about their borrowers’ characteristics was limited, their lending behaviour was influenced by the above noted illusions (marked as the lack of the first advantage on the list of Jappelli and Pagano), which considerably strengthened households’ incentive to over-borrow with the multiple lenders (reversing the fourth point on the list of Jappelli and Pagano).
The Hungarian authorities have acknowledged the increasing risk, trying to call the banks’ attention to its significance. First in 2004 and thereafter annualy, the MNB selected foreign currency lending as the biggest risk to the financial intermediation in its Reports on Financial Stability8, making serious efforts to regulate and restrict foreign currency lending. However, the regulatory bodies (esp. the Ministry of Finance) were not particularly eager to curb the credit boom.
In 2006, the HFSA issued a Recommendation ‘On the principles of retail crediting provision of preliminary advice to clients and consumer protection’9, followed by a series of other warning calls and recommendations by both the HFSA and MNB. However, market incentives and the above-discussed illusions were so strong that these warnings and recommendations have not led to any significant behavioural changes in banking until September 2008, when the failure of Lehmann quickly changed the rules of the game.
As a delayed response to the crisis, the banks did change their behaviour by sharply reducing their lending activity and by switching back from the foreign currency lending to forint lending (see Chart 4). With the lending activity falling back to about onequarter/ one-third of its pre-crisis levels, it seems to be a typical credit rationing behaviour, which is usually closely related to the lack of proper information.
Analysing the above chronicle, we call the Hungarian retail and, especially, mortgage markets ‘quasi-sub-prime’ (Király and Mérő, 2008) as, according to the generally accepted US definition, a weak and limited credit history results in the loan sub-prime market. However, at that time we have concluded that as the scarcity of information is coupled with a low delinquency rate, the picture about the real nature of Hungarian household lending could only have been made more accurate with the existence of a reliable country-wide credit information system.
Unfortunately, the crisis has confirmed our fears that Hungarian banks’ retail portfolios have a strong sub-prime character. This is clearly observable in the rising trend for the non-performing loans (see Table 1).
The portfolio of more than 90 days overdue loans more than doubled in the first year and a half of the crisis, with the value of defaulted mortgage loans almost tripling. Of course, not only a limited informational content, but also the quality characteristics of such loans might be a source of concern.
In Király and Mérő (2008), analysing the interest rate differentials between eurodenominated housing loans in the euro area and in Hungary during the pre-crisis period, we have also demonstrated the presence of the informational rent of Hungarian banks (the lack of the second advantage on the list of Jappelli and Pagano). This was possible due to very similar default rates on mortgage lending by the Hungarian and euro-zone banks. Our past analysis, based on the interest rate decomposition, can no longer be repeated – as, on the one hand, during the crisis the default rates became much higher than that of the euro-zone countries, and, on the other, interest rates became more strongly affected by the country’s increased risk premium. However, the crisis is very likely to have enforced the information rent, since the banks’ willingness to lend decreased significantly, concentrating only on the clients about whom they had extensive and exclusive information.


This paper summarised the history of a decade-long struggle for introducing a proper credit information system in Hungary. The lack of such a system has resulted in the presence of all the corresponding disadvantages as described in the academic literature. It also confers a quasi sub-prime character to the Hungarian retail lending portfolio. In the pre-Lehmann period, we had theoretically come to the conclusion that the Hungarian retail lending market was a quasi sub-prime one. Unfortunately, the high retail default rates of the post-Lehman period confirmed our fears. As a positive response to the crisis, we have observed higher risk awareness by both the banks and households, manifested in the shift from a foreign currency to the forint-denominated lending and to more conservative lending practices. However, lack of a reliable credit information system prevents the banks in having accurate information on their retail borrowers – which implies that the quasi sub-prime character of the Hungarian retail lending might remain for the times to come.
4 Éva Várhegyi, a well known banking scholar, headed the Committee. One of the authors (Julia Király) was a member of the Committee.
5 Letter of Intent of 29 October 2008.
6 One is a subsidiary of the multinational TransUnion‐CRIF Ltd, another is a new division of the above mentioned Hungarian BISZ Ltd.
7 Only contains credit granted by commercial banks (excluding lending by the cooperative banking system and finance companies).
8 See MNB (2004, 2005, 2006, 2007).
9 See HFSA (2006).

Tables at a glance

Table icon
Table 1

Figures at a glance

Figure 1 Figure 2 Figure 3 Figure 4
Figure 1 Figure 2 Figure 3 Figure 4