The banking sector has experienced drastic changes in the last few decades which were primarily driven by technological progress (DeYoung, et al., 2008) and bank consolidation(Petersen & Rajan, 2002). The consolidation has led to fewer branches and greater distancebetween banks and their borrowers (Petersen & Rajan, 2002, DeYoung, et al., 2008, Brevoort& Wolken, 2009). As a result, the availability of credit to small businesses decreased becausebanks could not assemble enough information over great distance to evaluate a borrower’sloan application (Craig & Hardee, 2007). Gathering information is a crucial part to assess aborrower’s creditworthiness. The collected information can be divided into hard and softinformation. Hard Information is defined as quantitative and is quickly reduced to numbersand convenient to transmit. Soft information is difficult to summarise in a numeric score andprimarily collected locally (Liberti & Petersen, 2017). The new lending technology (e.g. smallbusiness credit scoring (Berger & Udell, 2006)) which is based on hard information allowslarge banks to serve customers at a greater distance, increase credit availability for smallbusinesses and reduce costs (Petersen & Rajan, 2002, Berger, et al., 2005, Beck & Demirguc-Kunt, 2006, Brevoort & Wolken, 2009, Berger & Black, 2011). However, technological progresscan only partially substitute for the increase in distance because by replacing soft with hardinformation valuable information is not considered, which leads to the approval of bad loansand higher credit default rates (DeYoung, et al., 2008, Agarwal & Hauswald, 2010, Liberti &Petersen, 2017). To investigate the effect of technology on the importance of distance, thepaper is organised as follows. First, it will explain the capabilities of small business creditscoring (SBCS) and the positive effects associated with it; second, it examines the limits of thetechnology and shows why distance remains an important factor.
The most important technological innovation to impact the financing of small businesses isSBCS. There are several other technologies which have been established in the last years, butthe majority of the empirical studies only relate to small business credit scoring. Thetechnology is primarily based on hard information (e.g. financial statements) which isobtained through consumer and commercial credit bureaus. This information is then put intoa loan performance prediction model that estimates a score which displays thecreditworthiness of the borrower (Berger & Udell, 2006). One of the key findings with regardsto credit scoring was that a small business owner’s personal credit information is a predictorof his business creditworthiness (Berger, et al., 2005, DeYoung, et al., 2008). By using the
personal history of the owner rather than the business itself, lenders found a way to compensate for the missing soft information (Berger & Udell, 2006). Therefore, credit scoringis a valuable tool to help banks to evaluate the creditworthiness of a borrower who is locatedfurther away. The technology mitigates the information problems which were associated withlending to borrowers at a greater distance and led to increased credit availability for smallbusinesses (Frame, et al., 2001, Petersen & Rajan, 2002, Berger, et al., 2005, DeYoung, et al.,2008, Brevoort & Wolken, 2009, De la Torre, et al., 2010, Berger & Black, 2011, Berger, et al.,2011). Due to the greater availability of information and the processing of it by SBCS moredistant banks are now able to serve borrowers whom they would have refused in the past(Petersen & Rajan, 2002). As a result, SBCS made distance less critical to small businesslending and increased the quantity of loans granted. However, as will be argued later in thepaper, by the increased volume, the quality of loans decreases and the risk to approve badloans rises (DeYoung, et al., 2008).
By using hard information which does not have to be collected locally and is easy to transmit,the importance of distance decreased and large banks became able to lend at a greaterdistance. The substantial advantage of using hard information is that it can be easilytransmitted across distance without losing information (Liberti & Petersen, 2017).Furthermore, it is more durable, can easily be stored and can be interpreted without context,so it is possible to just pass it along to different people, and they can judge based on the data,without actually knowing the borrower (Stein, 2002). Therefore, hard information does notneed to be collected locally, and personal contact became a less significant part of thefinancing decision (Beck & Demirguc-Kunt, 2006, Agarwal & Hauswald, 2010). Although thecomplete information which is needed to evaluate a borrower’s creditworthiness is notavailable in “hard” form (e.g. numerical), some of the soft information can be “hardened”.This is particularly useful for large banks which were not able to process and judge softinformation due to their organisational complexity (Stein, 2002). Through hardening of softinformation and the use of SBCS also large banks became access to the small business lendingwhich resulted in a higher credit availability for small businesses and decreased geographicalimportance (Berger & DeYoung, 2006, Bartoli, et al., 2013).
The use of SBCS based on hard information is leading to reduced transportation costs for banks and cheaper loans for small businesses. Due to the technology, loan officers can cutdown on their monitoring, and the process gets less labour intensive (Petersen & Rajan,2002). The data collection, as well as the processing of it, can be delegated to workers withlower skills and partially automated by computers (Liberti & Petersen, 2017). Therefore, theadaption of SBCS which led to a more efficient process helped to reduce the agency costsassociated with lending at longer distances (Berger & DeYoung, 2006). Apart from increasedefficiency in the loan approval process, also the use of the owner’s rather than the company’scredit information significantly reduces costs (Beck & Demirguc-Kunt, 2006). The reducedcosts are particularly important for the area of small business lending. The size of loans givento small businesses are considered to be relatively low, so fix costs can make these loansexpensive. The fee which has to be paid is not depended on the size of the loan, so reducedcosts which result in lower fees will make loans considerably cheaper for small businesses(Petersen & Rajan, 2002, Liberti & Petersen, 2017). As a result, not only banks benefit fromreduced costs to originate a loan also the borrowers benefit from automatization due to lowerloan rates.
Even though credit scoring technology has brought some critical advances in the field of smallbusiness lending, the use of hard information can only partially substitute soft information.By merely replacing soft with hard information there is valuable information lost, so creditscoring technology evaluates a borrower’s creditworthiness solely on a set of reducedinformation which leads to less comprehensive results (Bartoli, et al., 2013, Liberti &Petersen, 2017). The technology only selects the part of the available information which iseasily reduced into numbers, so the other remaining large part of the high-quality informationis not taken into account (Agarwal & Hauswald, 2010). It does not include information aboutthe quality of the management or the prospects of the company which is gathered fromcommunication with suppliers, customers and neighbouring businesses (Berger & Udell,2006). This information is in particular important in an opaque market like small businesslending because there is a shortage of hard information and small businesses often only haveweak financial statements which are not audited (Breevort & Wolken, 2009, Cotugno, et al.,2013). Lending institutions are trying to compensate that by hardening soft information andmake it possible to use for the program in this way. The problem hereby is that this procedure erodes the quality of the information and therefore a lot of valuable information gets lost (Agarwal & Hauswald, 2010). Moreover, not all the information can be converted into a set of numbers. For example, the loan officer can use his discretion to more accurately evaluate the current creditworthiness of a borrower (Liberti & Petersen, 2017).
Small business lending crucially relies on firm-specific information which is collected locallyby loan officers. At this point, the importance of soft information cannot be dismissed by theuse of credit scoring technology. In particular, the area of small business lending tends to bean informationally sensitive market where cognitive processing and human judgment are vital(Agarwal & Hauswald, 2010, Knyazeva & Kyazeva, 2012). Banks which are located nearby theborrowers are better able to gather the soft information required to evaluate a smallbusiness. They have daily exposure to the local news, economy as well as they have a personalrelationship with the companies that give them the opportunity to gather soft information(Butler, 2008). In particular, relationship is a meaningful way to get the required informationwhich is not available in “hard” form. Therefore, loan officers with sufficient responsibility inthe loan approval process are needed. They build a business relationship and collect theinformation directly from the companies (Stein, 2002, Agarwal & Hauswald, 2010, Cotugno,et al., 2013, Liberti & Petersen, 2017). Soft information is also difficult to transfer, so even iflenders who are located further away want to buy soft information from a local loan officer,the transfer of the information would be a problem (Landier, et al., 2007, Alessandrini, et al.,2009). As a result, local banks have an advantage compared to non-local banks because theyhave access to soft information (Almazan, 2002, Degryse & Ongena, 2005, Butler, 2008,Agarwal & Hauswald, 2010, Liberti & Petersen, 2017). Lenders need soft information becausesmall businesses tend to be very intransparent. They only have weak financial statementswhich are not audited or available at all, and they offer collateral where the value is not easyto comprehend (Cotugno, et al., 2013). Therefore, assessing soft information is a crucial partto compensate for the weak quality of hard information. Empirical evidence shows thatdistance remains an important part of the loan approval process because the input of softinformation is needed to evaluate in particular opaque and low-quality borrowers like smallbusinesses (Brevoort & Wolken, 2009, Agarwal & Hauswald, 2010, Cotugno, et al., 2013, Iyer,et al., 2015, Liberti & Petersen, 2017).
The use of credit scoring leads, on average, to higher loan default rates because it increases lending to borrowers which are considered as riskier. As described in the beginning, theincreasing popularity of SBCS led banks to give credits to borrowers whom they would haverefused in the past (Petersen & Rajan, 2002, Berger, et al., 2005). By using credit scoringtechnology, there is less information collected, which leads to more approval errors andhigher credit default rates. As most lenders who use SBCS lend over greater distance, thisindicates that also the borrower-lender distance is positively related to the probability ofdefault. So, banks which are farther away have a higher chance of experience nonpayment(DeYoung, et al., 2008, Agarwal & Hauswald, 2010). However, Berger, Frame, & Miller (2005)also find statistically significant evidence that SBCS is positively associated with loan risk, butonly for loans below $100K. Loans in the range of $100K-$250K exhibit no substantialevidence of increased risk. But even though considering the risk for the loans under $100K,banks are willing to bear the higher default rates for the benefits (e.g. higher profit) of lendinga high quantity of loans using SBCS (DeYoung, et al., 2008). Nevertheless, DeYoung, Glennon,& Nigro (2008) also find that the negative impact of SBCS on loan default rates declines overtime because of the advances in technologies increases. This provides banks with betterabilities to evaluate the uncertainties of lending at long distances.
Loan rates are not only decreasing because of increased efficiency due to the use of SBCS, butthey are also decreasing because of increasing distance. The empirical literature providesevidence that loan rates tend to decrease with increasing distance between firms and theirlending banks, but increase with the distance between the lender and other competing banks(Degryse & Ongena, 2005, Butler, 2008, Agarwal & Hauswald, 2010). This indicates thatdistance still has important effects on loan pricing because otherwise every borrower shouldbe offered the same loan rate no matter how big the distance is. Degryse & Ongena (2005),who analysed around 18.000 loans in Belgium found significant evidence that borrowerswhich are very close to the lending bank are charged 14 basis points more, than borrowerswhich are located farther away. Although borrowers receive more affordable loans fromlenders who are located further away, they tend to decline these. Agarwal & Hauswald (2010)find highly significant evidence that more distant firms are more likely to refuse a bank loanoffer. This effect is even strengthened if the companies have a long lending businessrelationship with a bank (Agarwal & Hauswald, 2010).
- Quote paper
- Georg Aumüllner (Author), 2017, Effects of technology progress on banking, Munich, GRIN Verlag, https://www.grin.com/document/389076