Scholarly article on topic 'Social Lending and Its Risks'

Social Lending and Its Risks Academic research paper on "Economics and business"

CC BY-NC-ND
0
0
Share paper
OECD Field of science
Keywords
{"peer-to-peer financing" / "peer-to-peer lending platforms" / risk / "corporate entities" / "risk assessment"}

Abstract of research paper on Economics and business, author of scientific article — Martina Pokorná, Miroslav Sponer

Abstract Czech companies discover alternative non-banking financing of their business. Social lending known as peer to peer (P2P) financing has started to appear on the Czech market. P2P lending is a new platform of financial transactions that bypasses traditional intermediaries by directly connecting borrower and lenders. But there is an information asymmetry between lenders and borrowers to which online P2P lending platforms have to face. As many loans are not secured by collateral, the assessment of the borrower's creditworthiness is very important. The aim of this article is to define risks of investments to P2P loans and propose the elimination of these risks. We went through studies focused on default of corporate entities which were done in different countries and selected some of them. We have chosen these studies because their research is focused not only on financial indicators assessment but also on non-financial indicators assessment and use econometric models to measure probability of default. These studies also confirmed well known basic relations which should be taken into consideration by lenders – higher profitability, higher liquidity and higher volume of assets means lower risk of default, while higher indebtedness and higher leverage means higher risk of default.

Academic research paper on topic "Social Lending and Its Risks"

Available online at www.sciencedirect.com

ScienceDirect

Procedia - Social and Behavioral Sciences 220 (2016) 330 - 337

19th International Conference Enterprise and Competitive Environment 2016, ECE 2016, 10-11

March 2016, Brno, Czech Republic

Social lending and its risks

Martina Pokornaa*, Miroslav Sponera

aMasaryk University, Faculty of Economics and Administration, Department of Finance, Lipova 41a, 603 00 Brno, Czech Republic

Abstract

Czech companies discover alternative non-banking financing of their business. Social lending known as peer to peer (P2P) financing has started to appear on the Czech market. P2P lending is a new platform of financial transactions that bypasses traditional intermediaries by directly connecting borrower and lenders. But there is an information asymmetry between lenders and borrowers to which online P2P lending platforms have to face. As many loans are not secured by collateral, the assessment of the borrower's creditworthiness is very important. The aim of this article is to define risks of investments to P2P loans and propose the elimination of these risks. We went through studies focused on default of corporate entities which were done in different countries and selected some of them. We have chosen these studies because their research is focused not only on financial indicators assessment but also on non-financial indicators assessment and use econometric models to measure probability of default. These studies also confirmed well known basic relations which should be taken into consideration by lenders - higher profitability, higher liquidity and higher volume of assets means lower risk of default, while higher indebtedness and higher leverage means higher risk of default. © 2016 The Authors.Publishedby Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.Org/licenses/by-nc-nd/4.0/). Peer-review under responsibility of the organizing committee of ECE 2016

Keywords: peer-to-peer financing; peer-to-peer lending platforms, risk; corporate entities; risk assessment

World trend on lending markets is so called peer-to-peer platforms in several last years. P2P lending platforms bring to lending market a speed and efficiency thanks to rapid progress of modern IT technologies and easy accessibility of data and their processing. Platforms create their marketing image as progressive, effective and low cost alternatives of traditional financing.

First P2P lending platforms were established in the United Kingdom and in the United States. Their concept is to connect directly small investors with borrowers. There are also several P2P lending platforms in Czech Republic.

* Corresponding author.

E-mail address: m.pokec@seznam.cz

1877-0428 © 2016 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license

(http://creativecommons.org/licenses/by-nc-nd/4.0/).

Peer-review under responsibility of the organizing committee of ECE 2016

doi: 10.1016/j .sbspro. 2016.05.506

Seemingly attractive offer of progressive financial products should be critically considered. What are the differences in reality between services offered by a bank and P2P lending platforms' offers, what are advantages and risks.

Although P2P lending platforms are relatively new, many researches have studied the P2P lending market. Most of them have focused on finding factors that affect the probability of funding success and default rates in P2P lending. Avery et al. (2004) showed that a borrower's financial strength is crucial in ability to obtain secured and unsecured credit from financial institutions. Similarly, a borrower's financial strength plays an important role in the P2P lending market (Iyer et al. 2009, Herzenstein et al. 2008).

The aim of this article is to define risks of investments to P2P loans and propose the elimination of these risks. First part is focused on the benefits and risks of P2P lending. We introduce what P2P lending is, what is its progress and what are the advantages and disadvantages of online lending. We describe how the credit risk is evaluated by different investors in second part. The possibilities of credit risk elimination are proposed in the third part.

1. The benefits and risks of P2P lending platforms

P2P lending is a type of financial transactions concluded directly between individuals or between individuals and a company without the intermediation of a traditional financial institution. It has a short history but it has rapidly grown in last years. P2P lending platforms offer to borrower lower interest rates than banks usually offer. They are more flexible, more quick and transparent.

The first online P2P lending company was Zopa, launched in United Kingdom in 2005, next platforms were Prosper and Lending Club, established in the United States in 2006. P2P online exchanges are growing as alternative platforms to traditional savings and investments. Zopa is the UK's largest peer-to-peer lending service. They have lent £1.23 billion to over 150,000 people since 2005f. Prosper arranged loans in amount over 5 billion USD}. Lending Club arranged business loans in amount over 11 billion USD at the end of 2015 § and successfully entered into the stock exchange. Also British government invests through P2P lending platforms.

There are also several P2P lending platforms in Czech Republic. Intereson.cz was established in 2008 as the first P2P lending platform in Czech Republic. It is a part of Ferratum Group. Next P2P lending platform were Bankerat operating from 2010, pujcmefirme.cz (2014), Kreditni klub (2014) and Symcredit.cz operating from 2015. But only two of them - pujcmefirme.cz and Symcredit provide loans to corporate entities.

There are many advantages of P2P lending compared to loans transactions made through traditional lending institutions. The main advantage of P2P lending is that borrower can get a loan at a lower rate without collateral, while lender can obtain a higher return on his investments (Magee 2011). Despite the fact that high returns on investment from microfinance has been questioned, P2P lending has attracted a huge entry of investors who have been discouraged by the stock market returns and lower interest rates offered by banks (Brennan 2009). The Wall Street Journal has reported that the leading P2P platforms have provided investors with 10% or higher annual returns at a time of historically low interest rates (Haewon et al. 2012). The next big advantage of P2P lending platforms are transparent, flexible conditions, quick decisions and lower costs because operational costs of P2P lending platforms are not so administrative and hierarchic overloaded.

On the other hand, no investment is without risk, nor the investment to P2P loan is. In the online P2P lending market, the traditional role of credit risk assessment is left to individual lenders rather than financial institutions. Thus, there is possibility of misrepresentation for borrowers in terms of their creditworthiness (Haewon et al. 2012). Most of the requested loans are from borrowers which are not able to get a loan in a bank. If a borrower does not pay the loan back than is necessary to prosecute a claim with risk that the full amount of a loan will not be recover.

P2P loans are regulated only as so called "Providers of small sized payment services" in the Czech Republic. Because of that P2P lending platforms do not have an obligation to contribute to the "Fund of deposit's insurance" and investors do not have their investments insured.

f https://zopa.com ' https://prosper.com § https://lendingclub.com

There is a risk that P2P lending platform can go bankrupt. Most of P2P lending platforms established in Czech Republic do not make a profit till now, similarly as most of start-ups.

In the P2P lending market transaction costs are reduced by elimination of expensive intermediaries, but information asymmetry becomes serious problem. Most individual investors lack financial expertise and the lending experience takes place in a pseudonymous online environment (Klafft 2008). In this situation, social networks mitigate adverse selection (Lin et al. 2011).

It has been studied that players exhibit herding behaviors in online business when they face risk of uncertainty such as information asymmetry. Herding behavior describes many social and economic situations in which an individual's decision-making is highly influenced by the decisions of others. Duan et. al. (2009) suppose that herding behavior could be crucial especially in online business for two reasons. The first reason is information overload. There is an excessive amount of information on the internet, so users have difficulty to understand and use all the information. Doing what others do could be an efficient and rational way to make decision. The second reason is that people can easily observe others choices in online business. That is why most online e-commerce websites sort their products in the order of previous sales performance.

2. Credit risk evaluation

Every lender wants to be protected from the adverse consequences arising from credit risk. In order to credit risk would be as low as possible, lenders carry out an evaluation of the borrower. How the evaluation is done, how detailed it is, in which quality, it varies depending on financial knowledges of lenders.

2.1. Evaluation done by individuals

Investors have the worst position in P2P lending while evaluating borrowers. To efficiently allocate capital, the determination of acceptable interest rates must take the risk of default into account. How should individuals without knowing the borrower personally be capable of properly pricing default risk? Most of them are not financial expert, they can use only information provided by P2P platform - borrowers' credit scoring, financial data and information from the Internet. The existence of information asymmetries in the financial market is well known (Sufi 2007), but the information asymmetry between a borrower and potential lenders in the P2P lending market is even more acute. As Cheung (1989) has argued, the sustainability of any economic institution is subject to transaction costs associated with the organization. (Haewon et al. 2012)

When the lenders decide whether to invest their money in a loan request, they can verify the number of lenders who have already participated. If investors are influenced by the decisions of other investors (Devenow 1996, Welch 1992), this number is a kind of signal for lenders. In other words, an auction that already has many bidders may be more attractive to lenders considering investment. Eunkyoung et al. (2012) speculate that herding behavior is more common in this market due to the possibility of adverse selection and the limited institutional knowledge mentioned above when lenders face borrowers over the Internet.

2.2. P2P lending platform's evaluation

P2P lending platforms usually evaluate potential borrowers through rating agencies based on information from credit databases CRIF, CNCB and from the provided borrowers' financial statements. The crucial moment is the relation of the risk's evaluator to the borrower and to planned transaction. The bank evaluate the risk for itself and for its responsibility, P2P lending platforms evaluate risk seemingly independent. In reality, the commercial interest plays important role in the risk assessment - a margin from intermediation is paid to P2P lending platform when the loan transaction is concluded but the responsibility for default is on the side of investors. On the other side, the low default rate is good advertisement for P2P lending platforms.

For example the Prosper platform assigns each borrower a credit grade by using the financial documents borrowers provided. There are seven credit grades that vary from AA, signifying that the borrower is extremely low risk, to A,

B, C, D, E and HR denoting the borrower as of extremely high risk. Iyer et al. (2009) found that this credit score given to borrowers by Prosper is related to underlying creditworthiness and predicts the default rates.

Herzenstein et al. (2008) also used transaction data from Prosper and found that the probability of funding success of borrowers with AA or A is almost 40%, but the funding success rate of borrowers with HR is only 4%.

Similarly Symcredit in Czech Republic classifies credit scoring to borrowers. There are nine credit grades (A1, A2 - C3) with probability of default mentioned. Borrowers with credit scoring A1 have only +/- 0,06% of default, on the other side borrowers with credit scoring C3 - very high risk have +/- 7,50% of default.

2.3. Bank's evaluation

Client assessment in a bank includes an analysis of the legal situation of the loan's applicant and an analysis of the personal credibility of the applicant as a reliable business partner. A key part of the client creditworthiness verification is the client economic situation analysis. The analysis includes two related parts - the business and financial situation of the client. Bank lending decisions take into account status of a borrower as a manufacturer and trader. It analyzes the dynamic of firm's growth, position of a firm on the market in terms of its reputation, quality of products, whether it has diversified or highly specialized manufacturing program, its share on the market, the level of traded prices and costs in comparison with other competitors, etc.

The bank also assesses the branch position of the client and risks associated with the business. The current and future development in the relevant business is assessed, e.g. its increase, stability or decrease, a particular approach to the business of government (support or the effort to reduction).

After evaluating a client from a business point of view, more in-depth financial analysis of the basic financial statements - balance sheet, profit and loss and cash flow is assessed. Financial analysis measures received data among themselves and enhance their predictive ability against the basic financial statements, which provide current data, mostly in the form of state absolute quantities placed on a specific date, respectively flow values for a certain period.

Then the bank provides the analysis of financial ratios - indicators of profitability, activity, indebtedness and liquidity. Regarding the size of the calculated values, it is not possible to establish some firm recommended or even optimal values that have universal validity. Ratios do not constitute absolute precision scales for reporting characteristics of the enterprise, but they have largely probabilistic nature.

2.4. Other possibilities for credit risk analyzing

Other possibilities for analyzing and measuring the creditworthiness of corporate entities could be models built on conditional probability - logistic regression. It is about models of binary choice. These models are created from historical data. There are chosen variables which have the most information ability from original amount of variables. Prof. James A. Ohlson applied logit analysis on non-financial companies as the first in 1980. Even the first reported logistic regression prediction results had less predictive power than the ones reported in discriminant analysis studies, later on linear regression have shown that it has a powerful statistical approach for credit risk assessment. The above mentioned methods were criticized by Shumway (2001) or Hillegeist at al. (2004) and also Ohlson states in the conclusion of his research that account variables are not able to improve the model more and it is necessary to extend variables using market variables.

That is why new, innovative methods have appeared in last years, such as neural networks, decision trees, hazard models and support vector machines. Also recent literature indicates a strong interest in predicting credit risk applying methods from the field of artificial intelligence such as support vector machines or artificial neural networks (Mild et al. 2014).

Even with the existence of more sophisticated classification models for credit risk assessment the popularity and usage of logistic regression is growing due to its practicality and theoretical soundness. Aziz and Dar made comparison of these models in their study in 2006 with the aim to find the best one. The best results reached models of discriminant analysis and logit models according their study.

We went through studies focused on bankruptcy of corporate entities which were done in different countries and selected some of them where linear regression was used. Selected studies are from Finland (1999), France (2009), Malaysia (2015), Serbia (2013), United Kingdom (2014) and Czech Republic (2008).

Laitinen from Finland realized his study in 1999 and included 3200 companies from Finland with 5 years financial data. The model contents of 15 financial and non-financial variables which were classified as the background variables, payment history variables, responsible person variables, industry variables, financial variables about the company and about the group of companies. The results showed that financial ratios do not play as important role in credit rating as the background variables that describe payment history and the properties of the responsible persons. The shareholder's equity to total assets ratio seems to be the most important ratio.

Another study was done in France in 2009 by Psillaki et al. They included data of 4751 French companies operating between 2000 and 2004 in the textiles industry, wood and paper industries and computer activities and R&D industries. The model contents financial and non-financial indicators such as profitability, liquidity, leverage, turnover, collateral and growth opportunities. They compared results from the chosen industries and they found that non-financial performance indicators are useful ex-ante determinants of default. They showed that managerial inefficiencies are an important ex-ante indicator of the company's financial risk. Their results confirmed known relations - more efficient companies are less likely to default, profitability is important ex-ante indicator, more profitable companies are less likely to default, companies with more liquid assets have less chance to fail. The effects of leverage and growths opportunities varied across industries while collateral, the capital-turnover ratio and company's size generally had a negative relation with the probability of default.

The study realized in Malaysia in 2015 by Manab et al focused on non-financial indicator - earnings management. 30 companies - 15 healthy, 15 financially distressed were included with data 2006-2012. They constructed two models including four financial ratios - liquidity, profitability, productivity and leverage. First model was unadjusted and content above mentioned financial ratios, second one was adjusted so that the earnings management was subtracted from numerator and from denominator of particular book entries from which the ratios are calculated. For the unadjusted model, liquidity ratio and profitability ratio were significant while for the adjusted model, liquidity and productivity ratio were significant. Both models had similar accuracy rate in predicting distressed company, under Type I error. For Type II error the unadjusted model performed better than adjusted model. Including of earnings management indicator failed to increase the accuracy rate of predicting financial distress companies.

The authors from Serbia (Nikolic et. al. 2013) proposed in 2013 corporate entity credit scoring model capable of predicting probability of default in 1 year period. Dataset in this study consisted of 5 years financial statements dating from 2006 to 2011. The study took into account 7590 corporate entities. The list of 350 financial ratios has been constructed based on corporate financial statements and default event data. The study used brute force linear regression to come up with the most predictive credit scoring model. The weight of evidence data transformation technique has been applied in order to divide financial ratios into corresponding attributes and to eliminate problems connected to special values in financial ratios. Finally they proposed the credit scoring model that consist of 8 adjusted variables -debt ratio, leverage ratio, liquidity ratio, activity ratio, debt repaying capability (consist of two ratios), cash generating ratio and net sales growth ratio. They found that the most predictive variable is debt repaying capability which achieved information value of 0.923.

A study of around 340 000 international and 340 000 domestic SME companies was done in United Kingdom in 2014 by Gupta et al with the aim to compare the risk of international and domestic SMEs. Finally they used 9 financial and non-financial indicators such as retained earnings/total assets, cash/total assets, EBITDA/interest expenses, capital employed/total liabilities, log of current ratio, trade debtors/total assets, tax/total assets, trade creditors/total liabilities, intangible assets/total assets. They compared domestic and international SMEs separately but with the same ratios. They found that all the factors which affect the default probability of international SMEs are also highly significant in explaining the default probability of domestic SMEs. Furthermore, all the variables capturing the impact of exports on default probability of international firms are highly insignificant in the univariate analysis, thus contradicting the suggestion of Arslan and Karan (2009) to consider domestic and international firms separately while modelling their credit behaviors. The results confirm that the ratio intangible assets/total assets is highly significant in assessing credit risk for both domestic and international SMEs. They also found that non-financial factors may play an important role in understanding credit behaviors.

Iyer et al. (2009) differentiate this information between standard banking variables, which are based on hard, verified financial information from the credit report of the borrowers, and nonstandard variables. They indicated that econometric models outperform both credit scores in terms of the predictive ability regarding the risk of default. For

the econometric model, the benefit of using additional soft variables only marginally increases its performance. However, not all P2P lending markets have access to standard banking variables. The authors do not consider an econometric model that is solely based on soft variable in their paper.

There was also study in Czech Republic in 2008 done by Jakubik and Teply. They have tried to create aggregated model of creditworthiness for whole business sector in Czech Republic. They used 757 companies from which 151 were in bankrupt. They identified 22 variables from which 7 most predictable variables were selected based on logistic regression. The variables were liabilities/own capital, long-term payables/own capital, EBT/interest paid, EBIT/sales, (inventories/sales)x365, working capital/total assets, EAT/own capital. The model was statistically significant, Gini's coefficient was 80,41%. According their study the JT index was created. The probability of whole business sector default was from 2,6% to 2,9%.

Based on this analysis additional criteria for measuring credit risk could be established and it could be another supportive criterion for lenders in evaluating the creditworthiness of corporate entities. Moreover the above mentioned studies confirmed that logistic regression improved assessment done by discriminant analysis in all cases and also confirmed well known relations - higher profitability, higher liquidity and higher volume of assets means lower risk of default, while higher indebtedness, higher leverage means higher risk of default.

3. Possibilities for credit risk elimination

The majority of P2P lending platforms provide visitors with a database of historical loan project details such as interest rates and actual repayments. Lenders can make use of those sources of information to estimate a default risk. The proper fulfillment of this task can suffer from several limitations and biases. First, humans usually base their decisions on a few key figures than on a large variety of factors (Gigerenzer 2007, Shanteau 1992). Second, on P2P lending platforms, the majority of users do not consist of financial and/or mathematical experts but primarily of private persons that are not able to create complex models to base their investment strategies on.

Nevertheless lenders are able to decrease the risk by followings:

1. Invest only in borrowers without any delinquent accounts.

2. Invest your money according the risk which you are willing to accept. P2P loans are usually divided to several groups (A, B, C, D) according the borrower's scoring. Group A offers loans with lower return but with the lowest risk. On the other side, investors who would like to earn for example 11%, can invest to group D. But there is higher risk of borrower's default.

3. Invest several smaller amounts to several loans. For example 100 000 CZK could be split into ten different loans of ten different borrowers.

4. P2P lending platforms are not responsible for possible default and P2P lending is not regulated in Czech Republic. Nevertheless most of P2P lending platforms have their reserve funds from which possible investor's losses could be compensated. It holds in cases when P2P lending platform does not connect the lender and borrower directly.

5. Investors can invest not only to loans but also to P2P certificates. P2P certificates are offered for example by Symcredit in Czech Republic. P2P certificates are secured by a group of P2P loans issued by successful world P2P platforms. Investors are paid quarterly repayments of amount and interest based on fixed coupons. The maturity of the P2P certificates is from 3 to 5 years and offered interest rate varies from 6% to 8% p.a.

The credit risk could be decreased also by P2P lending platforms. They should provide customers clear and capital information. P2P lending platforms mostly intermediate borrowers' evaluation through rating agencies. There is a question, how the assessment is precise if the rating agency does not know a borrower personally. We consider possibilities to decrease credit risk by P2P lending platforms as follows:

6. Provide clear and capital information about a borrower.

7. Classify a borrower according probability of default, not according scoring.

8. Set key financial indicators with calculation when the concrete financial indicator reaches the possibility of default.

9. Include non-financial indicators into assessment.

10. Use econometric models which make the assessment more precise.

If investors and P2P platforms accept our suggestions then the investment risk to P2P lending could decrease and P2P lending become more attractive. It will mean higher profit for both sides - for investors and also for P2P lending platforms.

4. Conclusion

Whether you are a bank, financial institution or a private person, the lending of money is a risky investment by its nature. Lenders have to face different kind of risk including default risk, than the inaccurate assessment of credit risk can be a threat to the whole financial system as in the cases of the world financial crisis in 1929 and the subprime crisis in 2008.

The expansion of the internet has enabled a new form of matching supply and demand for capital - peer to peer lending platforms. Individuals and companies can present themselves and their planned projects and seek financing from private lenders there. Contrary to classical bank credits, borrowers and lenders are spread over the world without personal contact and the credit is usually not secured by collateral. The proper assessment of the credit default risk is the crucial task for lenders. Nevertheless lenders have to face also to other risks such as information asymmetry, herding behavior and adverse selection.

There are several possibilities on the side of investors as well as on the side of P2P lending platforms for credit risk elimination. Investors should invest only in borrowers with non-delinquents accounts and diversify their investments into several loans of several borrowers with different credit scoring. If P2P lending platforms provide clear and capital information about borrowers, classify a borrower according probability of default and not according scoring, if they set key financial indicators of borrower's financial situation including non-financial indicators and use econometric models as a support of traditional financial analysis, they could significantly help investors for right selection. If investors get enough clear and relevant information, the peer-to-peer lending will become more attractive and it could significantly grow.

References

Arslan, O., Karan, M.B. (2009) Credit risk and internationalization of SMEs. Journal of Bussiness, Economy and Management, 10, 361-368 Avery, R. B., Calem, P.S., Canner, G.B. (2004). Consumer credit scoring: do situational circumstances matter? Journal of Banking and Finance, 28, 4, 835-856.

Brennan, P-J. (2009). Peer-to-peer lending lures investors with 12% return (update 2). Bloomberg.

Cheung, S. (1989). Economic organization and transaction costs. The newPalgrave: Allocation, Information market. W.W. Norton & Company, New York, NY.

Devenow, A., Welch, I. (1996). Rational herding in financial economics. European Economic Review. 40, 603-615.

Duan W., Gu, B.,Whinston, A. (2009). Informational cascades and software adoption on the Internet: an imperical investigation. MIS Quarterly, 33, 1, 23-48.

Eunkyoung, L., Byungtae, L. (2012). Herding behavior in online P2P lending: An empirical investigation. Electronic Commerce Research and

Applications, Volume 11, 495-503. Gigerenzer Inc. (2010). Gut feelings: The intelligence of the unconscious. Viking Adult.

Gupta, J., Wilson, N., Greorgiou, A., Healy, J. (2014). The effect of internationalisation on modelling credit risk for SMEs: Evidence from UK

market. Int. Fin. Markets, Inst. and Money, pp. 397-413. Haewon, Y., Byungtae, L., Myungsin, Ch. (2012). From the wisdom of crowds to my own judgment in microfinance through online peer-to-peer

lending platforms. Electronic Commerce Research and Applications, Volume 11, Issue 5, 469-483. Herzenstein, M. Andrews, R. L., Dholakia, U. M., Liandres, E. (2008). The democratizations of personal consumer loans? Determinants of

success in online peer-to-peer lending communities. Working paper, SSRN. Available at ssrn.com/abstract=1147856. Hillegeist, S. A., Keating, K. E., Cram, D. P., Lundstedt, K. G. (2004). Assessing the probability of bankruptcy. Review of Accounting studies, vol. 9(1), pp. 5-34.

Iyer, R., Khwaja, A. I., Luttmer, E.F., Shue, K. (2009). Screening in new credit markets: can individual lenders infer borrower creditworthiness in

peer-to-peer-lending? Working paper, SSRN. Available at ssrn.com/abstract=1454976. Jakubik, P., Teply, P. (2008). The Prediction of Corporate Bankruptcy and Czech Economy's Financial Stability through Logit Analysis. IES

Working Paper 19/2008. Charles University. Klafft, M. (2008). Online peer-to-peer lending: A lenders'perspective. In H.R. Arabnia. & A. Bahrami (Eds.J. Proceedings of the international conference on E-learning, E-business, enterprise information systems, andE-government. EEE (pp. 371-375), Las Vegas.

Laitinen, E. K. (1999). Predicting a corporate credit analyst's risk estimate by logistic and linear models. International Review of Financial Analysis, pp. 97-121.

Lin, M., Prabhala, N., Viswanathan, S. (2011). Judging borrowers by the company they keep: friendship networks and information asymmetry in online peer-to-peer lending. Working paper, SSRN. Available at https://papers.ssrn.com.

Magee, J. (2011). Peer-to-peer lending in the United States: surviving after Dodd-Frank. North Carolina Banking Institue Journal, 139-174.

Manab, N. A., Theng, N. Y., Md-Rus, R. (2015). The Determinants of Credit Risk in Malaysia. Procedia — Social and Behavioral Sciences, pp. 301-308.

Mild, A., Waitz, M., Woeckl, J. (2015). How low can you go? - Overcoming the inability of lenders to set proper interest rates on unsecured peer-to-peer lending markets. Journal of Business Research, 68, 1291-1305.

Nikolic, N., Zarkic-Joksimovic, N., Stojanovski, D., Joksimovic, I. (2013). The application of brute force logistic regression to corporate credit scoring models: Evidence from Serbian financial statements. Expert Systems with Applications, pp. 5932-5944.

Ohlson, J. A. (1980). Financial ratios and the probabilistic prediction of bankruptcy. Journal of Accounting Research, pp. 109-131.

Psillaki, M., Tsolas, I. E., Margaritis, D. (2009). „Evaluation of credit risk based on firm performance. European Journal oof Operational Research, pp. 873-881.

Shanteau, J. (1992). How much information does an expert use? Is it relevant? Acta Psychologica, 81, 1.

Shumway, T. (2001). Forecasting bankruptcy more accurately: A simple hazard model. Journal of Business, vol. 74(1), pp. 101-124.

Sufi, A. (2007). Information asymmetry and financing arragement: evidence from syndicated loans. Journal of Finance, 62, 2, 393-410.

Welch, I. (1992). Sequential sales, learning and cascades. Journal of Finance. 47, 2, 695-732.