Source-linked AI summary
The Financing of Innovative SMEs: a multicriteria credit rating model
Silvia Angilella, Sebastiano Mazzù
TL;DR
Innovative SMEs face tighter credit constraints and often lack reliable financial information, creating a gap for judgemental credit-risk assessment based on qualitative criteria. The paper proposes ELECTRE-TRI within SMAA-TRI to assign SMEs to risk classes under parameter uncertainty. In the case study, uncertainty analysis leaves companies A and B unchanged but significantly lowers the assignments of companies C and D, revealing weaknesses and high financing risk for C and D.
Problem
Innovative SMEs face tighter credit restrictions and insufficient or unreliable financial data, while no judgemental multicriteria credit-scoring model for them had been identified.
Method
The paper combines ELECTRE-TRI with SMAA-TRI to assign innovative SMEs to risk classes using mainly soft information while accounting for parameter uncertainty and imprecision.
Results
Uncertainty analysis leaves companies A and B unchanged but significantly lowers the category assignments of companies C and D, revealing weaknesses and high financing risk.
Takeaways & Limitations
The approach provides a multicriteria framework for sorting innovative enterprises into risk classes using mainly soft information.
Takeaways & Limitations
The paper identifies detecting the critical criteria governing decisions as an unresolved area for further analysis.
Abstract
from arXiv · showhide
Small Medium-sized Enterprises (SMEs) face many obstacles when they try to access credit market. These obstacles are increased if the SMEs are innovative. In this case, financial data are insufficient or even not reliable. Thus, when building a judgemental rating model, mainly based on qualitative criteria (soft information), it is very important to finance SMEs' activities. Until now, there isn't a multicriteria credit risk model based on soft information for innovative SMEs. In this paper, we try to fill this gap by presenting a multicriteria credit risk model, specifically, ELECTRE-TRI. To obtain robust SMEs' assignments to the risk classes, a SMAA-TRI analysis is also implemented. In fact, SMAA-TRI incorporates ELECTRE-TRI by considering different sets of preference parameters with Monte Carlo simulations. Finally, we carry out some real case studies, with the aim of illustrating the multicriteria credit risk model proposed.
1 Introduction
Innovative SMEs face especially tight credit constraints because financial information may be insufficient or unreliable. The paper addresses this gap with a qualitative, multicriteria credit-rating model combining ELECTRE-TRI and SMAA-TRI, illustrated through Italian start-ups.
- Innovative SMEs face tighter credit-market restrictions than other SMEs, intensifying their financing difficulties.
- Banks use financial statements, collateral, hard-information scoring, and relationship lending, but innovative SMEs often lack reliable track records.
- The paper fills a stated gap by proposing a judgemental multicriteria credit-scoring model for innovative SMEs using qualitative information.
- The model combines ELECTRE-TRI with SMAA-TRI to evaluate risk and obtain more robust assignments under varying preference parameters and data uncertainty.
- A real case study applies the methodology to four Italian innovative start-ups using interviews with loan officers from a major Italian bank.
- The analysis also evaluates business-plan scenarios through NPV and financial ratios before incorporating those ratios into the multicriteria assessment.
2 A sorting model
The paper uses ELECTRE-TRI to sort alternatives into ordered risk categories through comparisons with reference profiles, then applies SMAA-TRI to assess assignment robustness under parameter uncertainty.
- ELECTRE-TRI: ELECTRE-TRI assigns innovative SMEs to predefined risk categories by comparing them with reference profiles delimiting those categories.
- ELECTRE-TRI: The concordance test measures weighted majority support for an outranking relation, while the discordance test captures minority criteria capable of vetoing it.
- ELECTRE-TRI: Preference and indifference thresholds represent, respectively, the smallest compatible preference difference and the largest difference preserving indifference.
- ELECTRE-TRI: An outranking relation is validated when its credibility exceeds the user-defined λ threshold, with λ constrained to [0.5, 1].
- Assignment: The study adopts the pessimistic assignment rule, which compares alternatives successively with profiles and assigns them to the corresponding category or the worst category.
- SMAA-TRI: SMAA-TRI incorporates uncertainty and imprecision in ELECTRE-TRI inputs by generating parameter sets through Monte Carlo simulations and evaluating category-assignment stability.
3 Description of the proposed model
The proposed model combines business-plan screening with qualitative and financial criteria, then uses ELECTRE-TRI to sort innovative SMEs into predefined risk classes structured around innovation-related risk areas.
- Preliminary financial analysis: The process begins with business-plan indicators such as NPV and uses base and worst scenarios to reject projects with NPV< 0.
- Preliminary financial analysis: Financial ratios are compared with sector quartiles, while NPV is estimated under each scenario as part of the preliminary analysis.
- Model rationale: Financial information alone is weakly significant for innovative SMEs when their specific risks are not analyzed.
- Model rationale: The model addresses risk sorting through multiple criteria, emphasizing non-financial qualitative criteria alongside financial information.
- Model structure: ELECTRE-TRI compares each SME with reference profiles to assign it to predefined risk categories through an interactive process involving decision makers.
- Criteria hierarchy: The selected financial criteria represent profitability, leverage, and liquidity, while innovation indicators include Intangible Assets/Fixed Assets and R&D/Sales.
- Criteria hierarchy: The criteria hierarchy covers development, technological, market, production, and financial risk areas, with sub-criteria organized beneath them.
4 Start-up case study
The case study applies the multicriteria model to four Italian innovative start-ups, combining business-plan financial analysis with qualitative credit-risk criteria and simulation-based assignments. Results produce risk classifications while exposing instability and potential weaknesses for Companies C and D.
- Case-study enterprises: The case study examines four Italian innovative SMEs using business plans and structured questionnaire data on company, market, network, supply-chain, ownership, and awards information.The firms are anonymized as Companies A–D and include biotechnology, digital communications, mechanical design, and nanomaterials start-ups.
- Financial analysis: Business-plan analysis uses base, cash-flow-reduced 20%, and cash-flow-reduced 40% scenarios to assess financial soundness.NPVs were computed under each scenario, and financial ratios were subsequently included in the multicriteria analysis.
- Financial analysis: All NPVs were positive for every enterprise under every scenario, so the projects were assessed as profitable before financial ratios were added.The analysis used a 7.93% risk-free rate and treated NPV as a screening tool for project profitability.
- Criteria construction: The model combines qualitative criteria, financial ratios benchmarked against sector quartiles, and expert-supported evaluations using binary or five-point codings.Quartiles from sector samples were used as ELECTRE-TRI limit profiles, while credit officers and experts evaluated criteria reflecting enterprise-specific risks.
- Risk assignments: Simulation-based assignments place Company A in C4, Company B in C5, Company C in C5, and Company D in C3 at the individual-DM level.Company A’s assignment is fully stable across DMs; Company B is assigned to C5 except by DM1, while Company C and Company D show some disagreement.
- Risk assignments: At group level, Companies C and D show instability, with Company C split between C4 and C5 and Company D split between C4 and C3.The authors use these results to identify weaknesses and risk factors affecting creditworthiness; the model may also involve type I and type II errors.
5 Discussion and managerial implications
The discussion shows how imprecise evaluations and simulation-based acceptability reveal instability and risk weaknesses, while highlighting important modeling choices and limitations for credit assessment.
- Model behavior: Category acceptabilities can behave non-monotonically: improving Company D’s evaluation may produce assignments to C3 or C5.For DM5, the corresponding acceptability indices are 63% for C3 and 47% for C5.
- Model behavior: Company D’s jump between non-consecutive classes depends on the λ threshold and identifies key competitors as a critical risk factor.Company D moves from class 3 to class 5 whenever λ is below 0.778, revealing enterprise weakness.
- Uncertainty analysis: Interval-based evaluations and a 20% prudential reduction of financial criteria were used to represent uncertainty in innovative-project data.The resulting category acceptabilities are reported in Table 10.
- Decision rules: With a probability threshold π4_i > 0.7, only Company A is assigned to class C4 in the case study.The threshold reflects the credit officer’s decisional perspective on the minimum probability required for assignment.
- Risk assessment: The uncertainty analysis leaves Companies A and B unchanged but lowers the categories of Companies C and D, revealing weaknesses in their credit profiles.These results are linked to the model’s non-monotonic behavior and support using the model as a risk assessment tool.
- Risk assessment: Companies C and D present higher financing risk because their sectors’ mortality rates exceed the national rate for Italian SMEs.The comparison uses sector mortality rates over 2008–2012 and is reported in Table 11.
- Model limitations: The model excludes veto criteria, although adding one could strongly alter assignments by preventing an enterprise from reaching a good risk class.For example, a missing unit pilot could force assignment to C1 despite good evaluations on most other criteria.
6 Conclusions
The paper presents an ELECTRE-TRI multicriteria model for sorting innovative enterprises into risk classes using soft and financial information. It identifies future work on decision-critical criteria and multicriteria models that represent interactions between criteria.
- Contribution: The approach sorts innovative enterprises into risk classes using innovation-risk indicators together with financial criteria.The innovation-risk structure includes development, production, market, and technological risks.
- Method: ELECTRE-TRI compares each enterprise with existing risk profiles through an outranking preference relation.The method is presented as suitable for the multidimensional and complex decision framework of financing innovation.
- Implementation: Credit officers and experts can define risk profiles and criteria weights and tune cutting, preference, indifference, and veto thresholds.These parameters affect the final ratings.
- Future research: Future work includes identifying the smallest criterion changes that would increase each enterprise’s acceptability for a better risk class.This defines the paper’s meaning of a “critical” criterion.
- Future research: Future work could also consider multicriteria models that represent interactions between criteria, including Choquet-integral approaches.The paper specifically notes the non-compensatory nature of ELECTRE-TRI and proposes hierarchical Choquet-integral methods for hierarchically structured criteria.
Appendix A
The appendix contains a scenario analysis with cash-flow reductions and a table comparing results with the sector.
- Scenario analysis: Table 12 is identified as a scenario analysis.The supplied passages list scenarios with cash flows lowered by 20% and 40%.
- Scenario analysis: The listed scenarios include cash flows lowered by 20%.This condition appears repeatedly among the supplied scenario entries.
- Scenario analysis: The listed scenarios include cash flows lowered by 40%.This condition appears repeatedly among the supplied scenario entries.
- Sector comparison: Table 13 is identified as comparisons with the sector.The supplied passages do not state the comparison outcomes.
Appendix B
The appendix describes Simos’ procedure for deriving criteria weights from decision-maker preference information and normalized weights.
- Weight assessment: Simos’ procedure uses z to express how many times the last criterion is more important than the first.The supplied definition treats z as a ratio of relative importance.
- Weight assessment: The procedure defines e′ r as the number of white cards between ranks r and r + 1.This variable represents the spacing between adjacent ranks in the preference ordering.
- Weight calculation: The non-normalized weight k(r) is computed within the Simos’ procedure, with e0 = 0.The appendix then distinguishes normalized weights from the sum of non-normalized weights.
- Weight results: Decision-maker preference information is reported in Table 14(a), while normalized weights are displayed in Table 14(b).Table 14 is identified as the Simos’ method with DM1’s set of weights and z = 8.