(AGENPARL) – Fri 07 November 2025 Mercati, infrastrutture, sistemi di pagamento
(Markets, Infrastructures, Payment Systems)
Do firms care about climate change risks?
Survey evidence from Italy
Number
November 2025
by Francesca Colletti, Francesco Columba, Manuel Cugliari, Alessandra Iannamorelli,
Paolo Parlamento and Laura Tozzi
Mercati, infrastrutture, sistemi di pagamento
(Markets, Infrastructures, Payment Systems)
Do firms care about climate change risks?
Survey evidence from Italy
by Francesca Colletti, Francesco Columba, Manuel Cugliari, Alessandra Iannamorelli,
Paolo Parlamento and Laura Tozzi
Number 70 – November 2025
The papers published in the ‘Markets, Infrastructures, Payment Systems’ series provide
information and analysis on aspects regarding the institutional duties of the Bank of
Italy in relation to the monitoring of financial markets and payment systems and the
development and management of the corresponding infrastructures in order to foster
a better understanding of these issues and stimulate discussion among institutions,
economic actors and citizens.
The views expressed in the papers are those of the authors and do not necessarily reflect
those of the Bank of Italy.
The series is available online at http://www.bancaditalia.it.
Printed copies can be requested from the Paolo Baffi Library:
Editorial Board: Stefano Siviero, Paolo Del Giovane, Massimo Doria,
Giuseppe Zingrillo, Paolo Libri, Guerino Ardizzi, Paolo Bramini, Francesco Columba,
Luca Filidi, Tiziana Pietraforte, Alfonso Puorro, Antonio Sparacino.
Secretariat: Yi Teresa Wu.
ISSN 2724-6418 (online)
ISSN 2724-640X (print)
Banca d’Italia
Via Nazionale, 91 – 00184 Rome – Italy
Designed and printing by the Printing and Publishing Division of the Bank of Italy
DO FIRMS CARE ABOUT CLIMATE CHANGE RISKS?
SURVEY EVIDENCE FROM ITALY
by Francesca Colletti,* Francesco Columba,** Manuel Cugliari,**
Alessandra Iannamorelli,** Paolo Parlamento** and Laura Tozzi***
Abstract
This paper presents the findings of a survey on climate change risk management by Italian non-financial
corporations, which was conducted by Banca d’Italia in 2024. The firm-level findings allow for a
more accurate assessment of climate-related risks and of their impact on creditworthiness. The survey
reveals widespread shortfalls in emission and physical risk management, transition planning, and
governance. Many firms that are not insured against physical risk tend to underestimate it. Moreover,
the approaches to climate change risk management exhibit marked heterogeneity, reflecting differences
in governance structure and across sectors and regions. Finally, the findings suggest that climate
sustainability commitments, if not accompanied by measurable progress, do not necessarily improve a
firm’s creditworthiness. The information processed for this work will be used to better integrate climate
change risks within Banca d’Italia’s In-house Credit Assessment System (ICAS).
JEL Classification: G32, Q51, C83.
Keywords: credit risk, climate change risks, survey methods.
Sintesi
Il lavoro presenta i risultati di un’indagine condotta dalla Banca d’Italia nel 2024 sulla gestione
dei rischi climatici da parte delle imprese non finanziarie italiane. I risultati a livello aziendale
consentono una valutazione più accurata dei rischi climatici e del loro effetto sul merito di credito.
L’analisi segnala carenze diffuse nella gestione delle emissioni e dei rischi fisici, nella pianificazione
della transizione e nella governance. Molte imprese prive di copertura assicurativa tendono a
sottovalutare il rischio fisico. Inoltre, gli approcci alla gestione del rischio climatico sono caratterizzati
da una pronunciata eterogeneità, riflettendo differenze nei modelli di governance, tra settori,
tra aree geografiche. Infine, i risultati suggeriscono che gli impegni ad accrescere la sostenibilità
climatica, se non accompagnati da progressi misurabili, non migliorano necessariamente il merito
di credito di un’impresa. Le informazioni elaborate in questo lavoro saranno utilizzate per rafforzare
l’integrazione dei rischi climatici nel sistema ICAS della Banca d’Italia.
Banca d’Italia, Turin Branch.
Banca d’Italia, Financial Risk Management Directorate.
Banca d’Italia, Florence Branch.
CONTENTS
1. Introduction
2. Structure
3. Sample
4. Respondents
5. Results
5.1 Climate Change Risk Governance
5.2 Emission Accounting
5.3 Investments in Emission Reduction
5.4 Investment Forecasts for the Energy Transition
5.5 Long-term Goals
5.6 Physical Risk
5.7 Insurance Coverage and Mitigation Strategies
5.8 Corporate governance section
6. Climate-Risk Adjusted PDs: ICAS granular data vs. sectoral approximation
6.1 Transition risk
6.2 Physical risk
7. Conclusions
References
Appendix
1. Introduction1
This paper presents the results of a survey devoted to climate change risk management practices by Italian
non-financial corporations (NFCs). The survey was conducted in 2024 as part of the ongoing efforts of
the In-house Credit Assessment System of Banca d’Italia (ICAS) to integrate climate change risks into its
evaluation process, using high-quality data.
On an annual basis ICAS produces ratings for about 4,000 firms, in two steps (Narizzano et al., 2024). In
the first step, a statistical model calculates the individual probability of default (PD) with a one-year
horizon, based on financial information and credit register data. In the second step, two analysts examine
further information and set the final rating. In complex cases, the rating is subject to the review of a rating
committee for the final decision. So far, in the second step, ICAS has applied a top-down approach for
climate risk assessment, using sector-level data on climate change risks (CCR), with limited use of firmlevel data.2
Climate transition and physical risks affect firms’ creditworthiness, via the impact on economic and
financial performance, and business model. The integration of climate-related risks into credit risk
assessment presents two main challenges. The first one pertains to methodological issues, such as those
related to the design of scenarios for climate change risks, the endogeneity of such risks, and the nonlinearity of climate effects.3 The second challenge is the scarcity of firm-level data on climate change
risks (Angelini, 2022b, 2023; Lavecchia et al., 2022).4 To address these challenges, the Eurosystem has
set minimum common standards for integrating climate change risks into the in-house credit assessment
systems managed by national central banks (NCBs).5 Such systems are a credit assessment source for
collateral in monetary policy operations; the largest commercial banks and the rating agencies also
maintain systems to that purpose (Giovannelli et al., 2020).
In 2024 Banca d’Italia carried out an experimental survey to collect information on firms’ greenhouse gas
emissions, energy consumption, financial impact of climate extreme events and qualitative information
We thank Paolo Del Giovane, Antonio Scalia and an anonymous referee for useful comments and suggestions.
Top-down approaches incorporate climate risk into credit assessment using sector-level or macroeconomic data,
applying broad adjustments across portfolios to capture systemic risk but with limited firm-specific detail. In contrast,
bottom-up approaches integrate firm-specific data, such as emissions and adaptation plans, into credit assessment for
tailored analysis of climate risks, though they require high-quality, granular data (Auria et al., 2021). A bottom-up
approach for ICAS climate risk methodology has been developed. However, it is currently being applied only to firms
participating in the EU Emissions Trading System (EU-ETS), due to the availability of firm-specific emissions data
(Cugliari et al., 2024).
The design of climate scenarios involves methodologies whose development is not settled yet, endogeneity refers to the
mutual influence between climate policies and economic behaviors, and non-linearity describes how climate effects can
rapidly escalate beyond certain thresholds (Monasterolo, 2020).
The sustainable data gap for analysing climate change and sustainable finance includes deficiencies in availability,
usability, access, and reliability of information (Network for Greening the Financial System – NGFS, 2021), also in Italy.
See ECB (2022). ICAS systems are currently used by the central banks of Austria, France, Germany, Greece, Italy,
Portugal, and Spain.
on corporate governance and organization. The objective of the survey was to enhance the evaluation of
the impact of CCR on credit risk and improve the accuracy of ICAS ratings by means of granular firmlevel data to be used in assessments for 2025.6 This approach overcomes the limitations of the use of
sectoral averages.
This paper presents the results of the survey and shows how they contribute to the evaluation of climaterelated risks. The survey identifies some shortfalls in Italian firms’ climate risk management, such as in
emission reporting and insufficient physical risk mitigation, highlighting significant differences with the
results expected on the basis of the available sectoral data.
According to the survey, in the manufacturing sector two thirds of the firms have governance bodies that
deal with CCR; in the services and agriculture sectors only half of the firms do. Less than half of the firms
track their emissions and almost all of them belong to sectors with high emissions; firms that do not track
emissions often deem them as being irrelevant or cite resource constraints. Stronger governance practices
and investments in emission reduction are more common in firms of higher credit worthiness; only a
minority of firms have formalized transition plans to achieve their climate goals.
The survey also reveals a disconnect between firms’ perceived exposure to physical risks and the potential
impact of extreme climate events. Most firms report low risks, despite one third of them being associated
with medium or high physical risk,7 yet they acknowledge potentially large disruptions to production
capacity, especially in the agriculture and manufacturing sectors. This evidence indicates that substantial
progress is still required in firm governance and data reporting in order for self-assessed climate risk
evaluations to be usefully integrated into credit risk assessment.
The survey responses illustrate the advantages of incorporating firm-level data into ICAS in the place of
currently available proxies. Regarding transition risk, PDs based on individual survey data significantly
differ from those estimated with sectoral averages. Sectoral data lead to an overestimation of PD compared
to the granularly estimates PD for 48 per cent of the firms in the survey and an underestimation for 52 per
cent of them. The mean and standard deviation of granular PDs are higher than for sectoral PDs.8 We
believe individual survey data also allow for a more accurate assessment of firm exposure to physical
risks than that based on sectoral data. Incorporating precise survey information on weather-related damage
and mitigation measures leads to an override of third-party scores for around 25 per cent of physical risk
assessments.
These results support the use of climate change survey data, for participating firms that responded to the
survey, for the purposes of ICAS, especially for firms that do not publish the non-financial statement, at
The sample of surveyed firms was targeted to maximize the amount of potential collateral to be assessed; as such, the
findings cannot be used for statistical inference on Italian NFCs as the sample is not random nor stratified.
The firms’ exposure to physical risk considers its headquarters and local units’ locations, with their relevance estimated
based on the number of employees.
The two figures are, respectively, 0.19 and 0.65 for granular PDs and, respectively, 0.03 and 0.07 for sectoral PDs.
least until more systematic firm-level information on CCR becomes available. In the coming years, in
fact, a number of initiatives are expected to increase the availability and quality of this type of information,
including the Corporate Sustainability Reporting Directive (CSRD)9 and, in Italy, the ‘MoF Platform on
Sustainable Finance’.10
The paper is organized as follows: section 2 outlines the structure of the survey; section 3 describes the
sample; section 4 provides details on the respondents; section 5 presents the main results; section 6
analyses the impact of granular data on climate risk-adjusted PDs; section 7 concludes.
2. Structure
The survey launched by Banca d’Italia intends to collect granular information from firms on CCR and
other qualitative data not otherwise available. Given the experimental nature of the survey, a limited
number of firms among those assessed by ICAS was contacted, selecting the most relevant ones in terms
of collateral. Participation in the survey was voluntary, but it was incentivized by supplying the
respondents with a feedback report comparing individual responses to those of a reference group.
The survey has two sections: (i) a first one on CCR; (ii) a second section on corporate governance and
other qualitative information. The climate section consists of 13 questions which span seven areas: i) the
existence of dedicated internal climate committees; ii) the monitoring of greenhouse gas (GHG)
emissions;11 iii) energy consumption data;12 iv) the assessment of physical risks and the firm’s ability to
manage them; v) the investments in energy transition to assess the firm’s commitment to reducing its
carbon emissions; vi) the emission targets; vii) the participation in the Emissions Trading Scheme (ETS).
The second section investigates corporate governance and organizational aspects that are relevant for
credit risk assessment of the firm. This section includes 12 questions that explore three areas: i) the firm’s
strategic positioning and competitive advantage; ii) the communication with stakeholders; iii) the
economic and financial outlook.
CSRD has updated and expanded the scope of the Non-Financial Reporting Directive (Directive 2014/95/EU, NFRD)
by requiring reporting for large, listed, companies from 2025, for large, unlisted, companies from 2026, for small and
medium-sized companies listed from 2027. Approximately 2,000 firms evaluated by ICAS should report such data by
2027. However, in February 2025 the EU Commission proposed to simplify sustainability reporting requirements
removing around 80 per cent of companies from the scope of CSRD. These changes are subject to approval by the
European Parliament and EU Member States.
The initiative is chaired by the Ministry of Economy and Finance (MoF) and involves the Ministry of Environment and
Energy Security, the Ministry of Enterprises and Made in Italy, Banca d’Italia, the Italian Companies and Stock Exchange
Commission (CONSOB), the Institute for the Supervision of Insurance (IVASS), and the Italian Pension Fund
Supervisory Authority (COVIP). Recently, the Sustainability Dialogue between SMEs and Banks was published by the
MoF Platform on Sustainable Finance. The document aims to support SMEs in gathering and producing information on
environmental, social, and governance (ESG) impacts, facilitating dialogue with banks on sustainability issues and
improving access to financing.
Including both direct and indirect emissions (Scope 1 and 2).
Renewable sources are distinguished from non-renewable sources to assess the carbon footprint.
3. Sample
To define the survey sample, we considered the 5,040 NFCs eligible for an ICAS full rating in 2024 (Fig.
1).13
Figure 1
NFCs eligible for ICAS full rating
(a) geographical area
(b) economic activity sector
10.9%
2.4% 4.1% 0.3%
4.3% 5.5%
16.2%
(c) CQS (1)
10.6%
41.6%
45.2%
34.1%
11.1%
43.7%
26.1%
31.3%
North West
Centre
North East
South & Islands
Manufacturing Services
Agriculture
Construction
Utilities
Mining
CQS 1&2
CQS 6
CQS 3
CQS 7
CQS 4
CQS 8
CQS 5
Source: Non-financial corporations eligible for ICAS full rating in 2024.
(1) Credit Quality Steps (CQS) are the credit risk categories defined by the Eurosystem to classify assets eligible as collateral in monetary policy operations. The scale
ensures the comparability of credit assessments across all accepted systems, such as ECAIs, IRB systems, and ICASs. CQS 1&2 and CQS 3 correspond approximately
to ECAI ratings from AAA to AA– and from A+ to A–, respectively.
The survey sample is designed to address specific information gaps for the ICAS expert assessment.
Hence the sample is not stratified and it should therefore not be used for statistical inference about the
population of all Italian NFCs.
We selected a candidate group of 1,102 firms from the ICAS sub-sample that we had initially identified,14
aiming to obtain responses from 680 firms, balancing the need for a sufficient number of valid responses
with the goal of containing the survey costs.15 The firms were selected with the following criteria: a) non-
Chosen according to criteria based on characteristics of the firm and of the associated loans eligible for use as collateral
from the 370,000 firms assessed with the statistical model.
Initially, 1,912 firms stood out as the most interesting for ICAS and selected according to purposive sampling, 871
were assessed relevant for climate change risks and 1,041 for governance and organizational aspects. The relevant
companies for climate change risks were identified as being ex ante exposed to transition or physical climate risks based
on Di Virgilio et al. (2024), external ratings, or third-party climate risk assessment services. Following the ineligibility
criteria of the American Association for Public Opinion Research (AAPOR, 2023; Smith T.W., 2023), for example, in
cases where the company could not be reached or was no longer operational, only 1,102 firms were included in the
candidate group (95 companies were also included in the Banca d’Italia – Survey of Industrial and Service Firms in 2023,
the so called Invind Survey).
The target included 540 significant firms in terms of climate change risk (with particular reference to exposure to
transition risk and participation in the EU ETS) and the remaining 140 chosen for qualitative significance (including those
with the highest statistical ratings or with a possible role as a parent company).
default status;16 b) medium or large size;17 c) relevance for monetary policy, based on the collateral
exposure of the lending banks. The survey was conducted by ICAS analysts in the regional branches
between February and June 2024.18
Figure 2 shows the distribution of the candidate NFCs by geographical area, sector, and credit quality step
(CQS). Additional details on the characteristics of the candidate group are provided in Table a2 in the
Appendix.
Figure 2
Candidate NFCs
(a) geographical area
(b) economic activity sector
1.3% 1.8% 0.3%
14.6%
(c) CQS(1)
5.9% 0.4%
44.1%
41.1%
49.9%
41.1%
27.5%
32.7%
North West
Center
North East
South & Islands
Manufacturing
Agriculture
Services
Construction
Utilities
Mining
CQS 1&2
CQS 6
CQS 3
CQS 7
CQS 4
CQS 8
CQS 5
CQS 9
Source: Non-financial corporations eligible for ICAS full rating in 2024.
(1) Includes CQS9 as some firms have gone into default since the start of the survey.
4. Respondents
The NFCs that responded to the survey are 577, more than half of the contacted firms, and not far from
the target of 680.19
The BI-ICAS default definition relies on Article 178 of the Capital Requirements Regulation (CRR), which sets forth
that a default occurs when a bank considers that the obligor is unlikely to pay its credit obligations or the obligor is past
due more than 90 days on any material credit obligation to the bank (Auria et al., 2021). A new harmonized definition of
‘fractional default’ has been introduced for evaluating the performance of ICAS models in the yearly Eurosystem CCAF
monitoring process (Narizzano et al., op. cit.).
According to Commission Recommendation 2003/361/EC, large companies are those with over 250 employees and
revenues above €50 million or assets over €43 million, while medium-sized firms have 50–250 employees and revenues
or assets between €10 million and €50 million.
Analysts reached out to companies by sending out questionnaires, followed by phone contact.
In the absence of statistical objectives and due to the non-priority nature of certain questions, even a single response is
considered valuable, as it provides otherwise unavailable information that ICAS analysts can use in creditworthiness
assessments (AAPOR, 2023; Ballin M. et al., 2000). The survey was submitted to 1,102 NFCs. The response rate can be
considered satisfactory as it is not far from that obtained in other Banca d’Italia surveys, in which there is often a
significant share of collaborative firms for several years (e.g. the participation rate was 64.2, 62.2 and 69.6 per cent,
respectively, for industrial, service and construction firms in the Banca d’Italia’s 2023 Survey of Industrial and Service
Firms, the so-called Invind).
The number of responses declined slightly as the questionnaire progressed beyond the first question, it
dropped after reaching 80 per cent of questions answered (Table 1). The most significant reduction
occurred in the climate section, where some questions required detailed responses.20
Table 1
Responses
Climate section
Corporate governance
section
Both sections
Only one section (1)
At least 1 answer
% of completed answers
Source: ICAS Survey.
(1) The column “only one section” reports the cases in which firms responded either to the climate section or to the corporate governance section,
Regarding the geographical distribution, 69 per cent of respondents are from the northern regions, a
slightly lower share compared to the ICAS population distribution (73 per cent). As for the distribution
across sectors, nearly half of the responses are from the manufacturing sector (48 per cent), followed by
services (40 per cent), in line with the distribution within the ICAS population. From a credit quality
perspective, 63 per cent of respondents are in the intermediate risk classes (Fig. 3),21 as in the ICAS subsample.
Figure 3
Respondents’ sample
(a) geographical area
(b) economic activity sector
(c) CQS(1)
2,6% 1,4% 0,2%
6.9% 0.3% 8.5%
35,4%
21,8%
48,0%
36.9%
40,0%
26.3%
33,3%
North West
North East
Center
South & Islands
Manufacturing
Services
Utilities
Agriculture
Construction
Mining
CQS 1&2
CQS 6
CQS 3
CQS 7
CQS 4
CQS 8
CQS 5
CQS 9
Source: ICAS Survey.
(1) Includes CQS9 as some firms have gone into default since the start of the survey.
Questions on energy consumption and emissions required detailed internal data that some firms, particularly those
without established environmental reporting, found difficult to provide.
CQS 3 and 4.
The share of respondents to the climate section that operate in the manufacturing sector is 57 per cent of
the section and the share in the higher credit quality steps is 35 per cent (CQS 1-3, Fig.4).
Figure 4
Climate Respondents’ sample
(a) geographical area
(b) economic sector
32.1%
(c) CQS(1)
0.6% 0.3%
12.6%
0.6% 6.6%
21.4%
28.9%
56.9%
12.9%
29.6%
36.8%
North West
Center
10.7%
North East
South & Islands
Manufacturing
Agriculture
Services
Construction
Utilities
Mining
CQS 1&2
CQS 6
CQS 3
CQS 7
20.4%
CQS 4
CQS 8
CQS 5
CQS 9
Source: ICAS Survey
(1) Includes CQS9 as some firms have gone into default since the start of the survey
Conversely, the firms that responded to the corporate governance section are more concentrated in the
services sector and more creditworthy. 22
5. Results
This section presents the main findings of the survey. The results provide insights into how Italian firms
approach climate-related risk governance, emission monitoring, and qualitative aspects relevant to the
ICAS framework.
5.1 Climate Change Risk Governance
The first section of the survey investigates how firms manage climate risk. 565 firms (98 per cent of the
respondents) provided an answer regarding the presence of a governing body responsible for climaterelated risk management.
For additional details on the respondents’ sample, see Tables a3-a5.
Figure 5
The results indicate in many cases the absence of
designated structures: 46 per cent of the NFCs report
Climate risk management
(percentage values)
no governance structure specifically tasked with
climate risk management (Fig. 5). This gap is
Sustainability Committee
prominent in the services and agricultural sectors,
Risk and Control Committee
where 58 and 60 per cent of firms, respectively, lack a
Nominations Committee
designated governance body; the share drops to 35 per
Other board committee
cent in manufacturing. A minority of firms (25 per
Sole administrator or Board of
Directors (BoD)
cent) indicate that they have structures exclusively
Dedicated figures external to the
devoted to climate risk management (sustainability
No dedicated organ
committee or dedicated figures external to the Board
Directors,
BoD),
frequently
Source: ICAS Survey
manufacturing sector. Additionally, in nearly one-third of the firms, climate-related responsibilities are
integrated into existing governance bodies (risk and control committee, nominations committee, other
board committee, sole administrator or BoD).
The absence of designated bodies is less frequent among firms with higher credit quality scores:
only 37 per cent of the NFCs in Credit Quality Step (CQS) 1&2 and 39 per cent in CQS 3 report no
specific climate governance roles. In contrast, the share rises to 73 per cent in CQS 6 and 62 per cent in
CQS 7 and 8. This pattern should not be interpreted as an indicator of a causal relation: lower attention to
climate risks does not necessarily lead to lower credit scores. Causality, if any, may also be running in the
opposite direction (reverse causality): NFCs with better credit profiles may naturally adopt stronger
governance practices, including those related to climate risks, due to a broader attention to organizational
resilience.
5.2 Emission Accounting
Another section of the survey examines the tracking of direct and indirect greenhouse gas emissions
(Scope 1 and Scope 2 emissions, respectively). Of the 554 respondents to this section of the questionnaire,
only 44 per cent state that they track emissions and only one tenth of the remaining firms plan to begin
the tracking; awareness and prioritization of climate risks thus show ample room for improvement.
Emission accounting is more widespread in the manufacturing sector, where the majority of firms
(54 per cent) monitor their emissions.
Figure 6
Reasons for not reporting emissions
(percentage values)
Of the firms that do not track emissions, 40 per cent
cite “lack of relevance or significance” as the primary
reason and 14 per cent the lack of skills or resources
(Fig. 6). Among the firms that do not track emissions
Lack of expertise for calculation
only three per cent belong to low-emission sectors,23
81 per cent belong to medium-emission sectors and 16
Lack of resources to dedicate
per cent belong to high-emission sectors.24 The firms
Lack of relevance / materiality
in sectors with high emissions report that the lack of
tracking is due in 22 per cent of the cases to their own
Other reasons
emissions being not significant25 and in 24 per cent of
cases to the lack of skills or resources.
Source: ICAS Survey.
5.3 Investments in Emission Reduction
The survey also examines firms’ investments towards reducing their greenhouse gas emissions (“ecosustainable investments”). Among the 524 respondents to this question, 40 per cent report having made
such investments in 2023. Investments in emission reduction are more common among firms with
higher credit quality scores: 55 per cent of the respondent NFCs with CQS 1&2 and 46 per cent of those
with CQS 3 invest in such initiatives. This finding may suggest a relation between credit quality and the
(willingness and) capacity to allocate resources to emission reduction, though it is unclear whether
stronger credit profiles act as an enabling factor for such investments or whether the underlying
organizational strength supports both creditworthiness and sustainability.
In addition, the survey revealed that large firms are more financially committed in reducing their
environmental impact; about 80 per cent of the companies that made investments towards reducing their
greenhouse gas emissions in 2023 were large. This finding aligns with the ECB’s Survey on the Access
to Finance of Enterprises (SAFE; Ferrando et al., 2023).
On average, firms indicate 1.4 financing instruments. Self-financing is the most common funding
source for investments towards reducing own emissions, reported by 75 per cent of firms, followed by
bank loans, indicated by 45 per cent of respondents. Public funding is used by only 11 per cent of the
firms; this share rises in Southern Italy and the Islands to 27 per cent of the respondents. Firms with
better CQS scores favour self-financing (91 per cent); those with CQS 3 to 5 rely more on bank
loans. Financing instruments such as bonds and equity are less common yet finance substantial portions
The classification of the sectors with low, medium and high emissions is described in Di Virgilio et al. (2024).
Of these firms in high emission sectors, 34 per cent belong to the land transportation sector, 26 per cent to the metal
sector, 16 per cent to the chemical sector.
This may be the case, for example, for firms operating in paper production or in the metallurgical industry, generally
considered as high-emissions entities.
of investments when used. Specifically, bonds fund on average 39 per cent of investment, equity 66 per
cent and inter-firm financing over 70 per cent (Table 2). Only 11 per cent of firms resorts to public
funding, which on average covers only 37 per cent of eco-sustainable investments. These results confirm
the findings for the euro area from SAFE, which identifies public subsidies to climate-related investment
as insufficient.
Table 2
Distribution of financing instruments for eco-sustainable investments
(percentage values)
Type of financing
Firms (1)
Self-financing
Average share of investment (2)
Interfirm financing
Banks and other financial intermediaries
Issuance of bonds and similar instruments
Equity
Public funding
Other
Source: ICAS survey.
(1) Percentage of firms using the specific financing instrument; – (2) Average share of investments financed by this financing instruments, calculated
only for firms using it.
Figure 7
Reasons for no eco-sustainable investments
(percentage values)
For firms that do not invest in the reduction of
emissions, 56 per cent cite “other reasons”,
including ongoing planning due to regulatory
uncertainty. This is consistent with the idea that a
Not feasible due to the production
structure
clear regulatory framework could stimulate
Not feasible in general
investment. Additionally, 16 per cent of the
respondents report that further investments are not
Difficulty in obtaining financing
economically viable due to the nature of their
Investments already completed
production processes, highlighting sector-specific
challenges (Fig. 7).
Other reasons
Source: ICAS Survey
10% 20% 30% 40% 50% 60%
5.4 Investment Forecasts for the Energy Transition
In terms of economic outlook, 76 per cent of
responding firms provide forecasts for own investments in the energy transition. Among these, 43 per
cent plan no new investments, 25 per cent plan to increase their investment flows, and 24 per cent expect
to maintain current flows. Only a small portion foresees reducing their annual investments, while 19
per cent of the companies that reported no investments in the previous year foresee them in the future.
5.5 Long-term Goals
Among the 519 respondents to this question, 43 per cent have set emission reduction targets for the
Figure 8
next five years. These targets are more widespread in
Emissions and non-renewable energy reduction targets
the construction and manufacturing sectors (63 and 51
(percentage values)
per cent, respectively) and less frequent in the services
sector (33 per cent). Over two-thirds of firms in CQS
1&2 have set targets, while this share declines
Scope 1 Emission Reduction
significantly for firms in lower CQS. For the firms
with established targets, the majority (over 80 per
Scope 2 Emission Reduction
cent, Fig. 8) expect reducing Scope 1 emissions
relative to 2023. Nearly three-fifths of respondents
Reduction of Non-Renewable
Energy Use
report having no emission reduction targets; a similar
Source: ICAS Survey
share has not made any related investments.
Just over 60 per cent of the respondents expect a
reduction in Scope 2 emissions or in the use of non-renewable energy sources.
Among firms that have set emission reduction targets
for the next five years, however, only a limited
Figure 9
Perceived exposure to climate risk
(percentage values)
number of firms have formalized a plan to achieve
these goals, more often in the manufacturing sector.
No clear relation emerges between having defined
a transition plan and the firm CQS score. Many
1 = not at risk
2 = low risk
firms appear to lack the governance and resources
necessary to translate the objectives into actions,
regardless of the credit standing; more than half of the
firms that have not defined a plan in fact have not
identified a body responsible for following up on
3 = medium risk
4 = high risk
5 = very high risk
sustainability issues.
Source: ICAS Survey
5.6 Physical Risk
Among the 541 respondents to this question, 25 per cent of firms report to have suffered direct damages
from extreme weather events over the past five years, with a higher incidence in the North-East and NorthWest areas. However, the impact on productive capacity is low, with more significant effects in the
agriculture and utilities sectors.
In terms of perceived exposure to physical climate risk, 70 per cent of the firms indicate a low or
zero risk level; of these 68 per cent, according to data based on ISPRA maps, appear to be in areas
with low physical risk, 19 per cent in medium risk areas and 13 per cent in high risk areas.26 Only
one per cent of firms, all in northern Italy, assess their exposure as very high (Fig. 9). To evaluate
the accuracy of this risk perception, self-assessed risk scores (on a 1-to-5 scale) are compared with those
from an external provider. Across the 432 respondents to this question, 29 per cent show alignment
between internal and external scores; 35 per cent of firms underestimate the risk compared with the
external provider, while 36 per cent overestimate risk. This symmetric distribution reveals no systematic
bias, but rather a prevailing misalignment between perceived and externally assessed exposure. The
divergence is particularly evident among firms operating in areas with a high physical risk: although
nearly one-fifth of the respondents operate in such areas, more than two-thirds of them perceive their
exposure as minimal, and the majority of those acknowledging some exposure report only limited
expected impacts on production capacity.27 These findings suggest that firms’ perception of climaterelated physical risk may be disconnected from objective indicators of exposure and vulnerability,
although we are not in a position to verify this possibility.
Despite the moderate risk perception reported by firms, the potential impact of extreme climate
events on productive capacity is assessed as more substantial: based on the 279 responses received to
the dedicated question, the average estimated impact is 19 per cent. This apparent contradiction may stem
from the distinction between perceived probability (“exposure”) and perceived severity (“vulnerability”):
firms may consider the occurrence of extreme events unlikely, yet still recognize their potentially
significant consequences on operations if such events were to materialize. These estimates of the potential
impact of extreme climate events are significantly higher than the damages reported over the past five
years, likely reflecting either an underestimation of past losses or expectations of worsening climatic
phenomena.
The discrepancy between the low perceived risk and the high expected impact may reflect the
coexistence of limited recent experience and growing awareness of the potential consequences of
extreme events. This could be due to heightened salience of recent climate disasters without
excluding a dynamic adjustment of expectations over time. Although the recent impacts on productive
capacity have been limited, firms may be factoring in the growing intensity of climate-related risks. This
apparent contradiction may reflect cognitive and structural aspects of risk perception. Firms may perceive
low risk based on the limited frequency of past events, influenced by short institutional memory and a
lack of recent direct experience. At the same time, recent high-impact disasters – widely reported in the
news and socially salient – may amplify their perception of potential severity (Kunreuther et al., 2014).
The firms answer for the group, if existing, to which they belong. We compare the survey responses with data based
on ISPRA risk maps.
The external provider assigns a five-tier score based on the geographical location of the company’s headquarters and
any secondary sites (for the latter, information on the number of assigned employees, where available, is used to estimate
their relative importance). For the purposes of this paper, firms with a score in the two lowest tiers are classified as ‘high
physical risk’.
From a modelling perspective, this aligns with the standard separation between the probability of
occurrence and the conditional severity distribution: climate-related physical risks are often characterised
by low expected frequency, but potentially extreme consequences, consistent with fat-tailed loss
distributions (Batteson et al., 2014).28
5.7 Insurance Coverage and Mitigation Strategies
The survey indicates that 73 per cent of firms are insured against physical risks, with higher coverage
among manufacturing and services (81 and 67 per
cent, respectively); most of these firms are large ones
Figure 10
Reasons for not having an insurance coverage
(percentage value)
(76 per cent). These results are in line with the
findings of Banca d’Italia’s 2021 Survey of Industrial
and Service Firms (so-called Invind), which showed
that 68 per cent of firms had bought insurance
Unsustainable economic cost
Sustainable cost but excessive
insurace premium
coverage at the time.29 However, the reasons for the
Risk is not relevant
lack of insurance coverage differ across the two
Lack of information on insurance
products
surveys. In the ICAS survey, about one-third of Lack of trust in insurance companies
respondents cite non-economic factors, such as the
Other reasons
perceived insignificance of risk, while Invind
10% 20% 30% 40% 50% 60%
participants primarily flag high insurance costs and Source: ICAS Survey
insufficient information. These differences may reflect the characteristics of the respective samples: the
higher share of smaller firms in the Invind sample may explain their greater focus on cost-related barriers.
Participation in the insurance market is lower for smaller firms and for those located in the South and
Islands consistently with previous analyses (Angelini, 2022a; Gallo et al., 2022; Frigo and Venturini,
2024). Insurance coverage is instead more widespread among firms with higher CQS ratings
(coverage rates between 73 and 81 per cent), compared to firms with lower CQS scores, which display
coverage rates between 41 per cent (CQS 7) and 63 per cent (CQS 6).
Among the firms without insurance coverage, 58 per cent indicate that they consider climate risk
irrelevant for their operations (Fig. 10). Common strategies include investment in safety measures, the
diversification of production and storage sites, general insurance, disaster recovery plans, and relocating
operations to areas with lower exposure to climate risk. Italian firms will nevertheless have to gradually
The Third UK Climate Change Risk Assessment Technical Report (CCRA, 2021) and Rising et al. (2022) report that
risk perception often trails empirical risk indicators for firms with limited resources devoted to resilience.
The same survey also found that damages from climate events more than doubled between 2016 and 2021, while
insurance adoption remained largely unchanged (Gallo et al., cit.).
adapt to the national regulations concerning compulsory insurance against damages resulting from
catastrophic events.30
5.8 Corporate governance section
This section presents the findings from the survey’s second section, which does not investigate climate
change risks, but contributes to hone the qualitative assessment performed by ICAS analysts of key areas
of firms’ governance that affect its creditworthiness and on which alternative data are not available. The
questions in this section investigate corporate practices in market monitoring, resource allocation for risk
management, and formal control functions. In particular, we investigate the scope of governance
structures and strategic planning across firms, including plans to respond to sectoral challenges and
emerging risks.
Regarding market monitoring, 79 per cent of 535
respondents state that they monitor trends and
Figure 11
Corporate control functions
(percentage values)
potential opportunities or threats in their sector of
activity. Only one third allocate specific resources to
this function, with higher CQS-rated firms doing so
more frequently than lower-rated ones. This practice
Risk Management
Compliance
is observed more frequently in the manufacturing
sector (42 per cent) than in the services sector (30 per
Internal Audit
Supervisory Body pursuant to
Legislative Decree 231/01
cent).
For formal control functions, 78 per cent of the
None of the above
respondents report to have a Supervisory Body
under Italian Legislative Decree 231/2001. Other Source: ICAS Survey
control functions including internal audit, compliance and risk management, are less widespread (Fig.
Regarding social impact reporting, 93 per cent of firms responded, with 52 per cent stating that they
do not produce dedicated documentation, such as sustainability or social responsibility reports.
Reporting is more common in sectors like utilities and construction, where 55 and 63 per cent of firms,
respectively, engage in such reporting.
Insurance coverage for asset damage caused by natural disasters was made mandatory in Italy (Law 213/2023). The
Italian Budget Law for 2024 has envisaged a gradual entry into force of the obligation for Italian firms, differentiating
the recipients between large, medium, small and micro enterprises (by 31 March, 1 October and 31 December 2025,
respectively).
6. Climate-Risk Adjusted PDs: ICAS granular data vs. sectoral approximation
The ICAS climate risk survey aims at collecting data for enhancing the precision of climate risk
assessment. To check whether firm-specific transition risk (TR) information improves the accuracy of
credit risk assessment, the survey results are compared with the sectoral estimates31 of emission data used
so far in credit risk methodologies. For physical risk, the scores provided by an external provider are
refined using the survey data described in previous sections.
6.1 Transition risk
Firstly, we perform a comparison of emission data. For each firm included in the survey, we obtain
greenhouse gas emissions by aggregating data reported by firms for the Survey for Scope 1, and Scope 2
(either location-based or market-based) emissions.32 When Scope 2 market-based data are available, we
aggregate Scope 1 with Scope 2 market-based data to reflect the energy procurement choices of the firm,
such as renewable energy contracts or supplier-specific emission factors. This method prioritises firmspecific information, recognising that Scope 2 market-based data offer a more accurate representation of
a firm carbon footprint than Scope 2 location-based data. For firms that do not provide data on direct
Scope 2 emissions, energy consumption is converted into emissions using the ISPRA emission factors
(ISPRA, 2023),33 which provide standard values for key energy sources.34 This process ensures
consistency and comparability across firms, reflecting Italy’s specific energy mix and decarbonization
trends as documented by ISPRA. The resulting emission dataset covers 432 firms, representing 11.5 per
cent of Italy’s total emissions in 2022.
Among the 83 firms included in the survey declaring participation in the EU-ETS, 67 also reported Scope
1 emissions in the survey. For 54 of them, verified emissions from the EU-ETS registry were retrieved,
showing a strong correlation (0.87) and an average standard deviation below 9 per cent. This alignment
underscores the accuracy of self-reported data and confirms the survey’s potential as a validation tool for
Sectoral estimates refer to emission values and transition risk indicators derived from industry-level averages rather
than firm-level data. These estimates are typically obtained by imputing energy use and emissions based on the firm’s
sector classification (e.g. NACE codes), employment figures, and standard energy intensities.
Scope 1 emissions refer to direct greenhouse gas emissions from sources that are owned or controlled by the company,
such as emissions from fuel combustion in company-operated facilities or vehicles. Scope 2 emissions are indirect
emissions associated with the consumption of purchased electricity, heat, or steam, and depend on how the electricity is
generated and procured. Firms provide Scope 2 data in two distinct categories: location-based and market-based,
reflecting the different methodologies for accounting emissions from purchased electricity. Location-based data represent
emissions calculated using the average emission intensity of the grid where the electricity is consumed, while marketbased data are derived from supplier-specific emission factors or contractual instruments, such as renewable energy
certificates.
ISPRA (Istituto Superiore per la Protezione e la Ricerca Ambientale) is the Italian Institute for Environmental Protection
and Research. It operates within the framework of the National System for Environmental Protection (SNPA) and is
responsible for conducting research, providing technical-scientific support, and developing indicators and reports related
to environmental protection and sustainability in Italy.
For instance, emissions from natural gas are calculated using a factor of 0.001927 tCO₂e per standard cubic meter
(Sm³), while market-based electricity emissions are computed using a factor of 0.4491 tCO₂e per megawatt-hour (MWh).
information independently collected by ICAS, despite minor discrepancies due to reference years or
reporting boundaries.
The comparison over the whole set of 432 firms reveals significant discrepancies between individual and
sectoral emission levels (Fig. 12). The absolute difference between emissions levels exceeds 10 per cent
in 85.4 per cent of cases, highlighting substantial discrepancies between firm-level and sectoral estimates.
A visual inspection of the scatter plot in the left panel suggests a systematic prevalence of firms whose
granular emissions exceed the corresponding sectoral estimates, as evidenced by the dense concentration
of points above the 45-degree line. The right panel complements this evidence by showing the distribution
of the logarithmic difference between granular and sectoral emissions. The asymmetry in the distribution
– particularly the predominance of green points – further confirms that firm-level data often report higher
emissions than sectoral averages.
Figure 12
Individual vs sectoral emission levels
(logarithmic scale)
Source: authors’ calculations based on sectoral or individual emission data.
Note: the left panel shows individual and sectoral emissions, while the right panel illustrates the dispersion of the log-difference between
individual and sectoral emissions across our 432 firms.
Secondly, we integrate these data into the ICAS framework to assess their impact on credit risk metrics.
We incorporate individual emission data into the Banca d’Italia’s climate stress-testing framework,
building on Faiella et al. (2021) and Di Virgilio et al. (2024). This methodology relies on a tool that
embeds the financial risk associated with climate policies into the ICAS methodology. Specifically,
emission data are fed into the ICAS stress-testing framework to simulate the impact of TR on the PD of
Italian firms. By computing energy expenditures and working out their impact over a range of financial
statement items, the methodology allows for the calculation of key financial variables used as inputs to
the ICAS statistical model. The mean and standard deviation of granular PDs (respectively, 0.19 and 0.65,
Table 3) are higher than the equivalent figures for sectoral PDs (respectively 0.03 and 0.07)35, as using
firm-level data naturally introduces greater variability compared to assigning the same sectoral value to
all firms. While this result reflects the shift from sectoral averages to firm-level information, it is
methodologically relevant: the increased dispersion and higher average PD highlight the capacity of
granular data to reveal firm-specific vulnerabilities and risks that are masked when sectoral proxies are
used. This finer granularity improves the ability of the ICAS framework to capture the heterogeneity of
transition risks across firms, which is critical for a forward-looking credit risk assessment.
Table 3
Sectoral vs Individual PDs
(percentage values)
Method
Std. Dev.
Sectoral
Individual
Source: authors’ PDs calculations based on sectoral or individual emission data.
The impact of using firm-level versus sectoral data becomes even more evident when comparing the yearon-year change in TR-adjusted PDs derived from the two datasets (Fig. 13).36
Figure 13
Change in 1-year credit ratings
(percentage values)
Source: authors’ TR-adjusted PDs calculations based on sectoral or individual emission data.
Both PDs are computed at the individual firm level; the former (granular PDs) use firm-specific data collected via the
ICAS survey, while the latter (sectoral PDs) are based on sectoral averages for climate change risk variables.
The individual emissions exhibit a wider dispersion and far more outliers, capturing higher variability and extreme
values in the transition-risk adjusted PDs.
The diagonal line represents perfect alignment between the two estimates, while deviations indicate firms
whose 1-year PDs differ depending on whether granular or sectoral emission data are used. Out of the 172
firms for which complete Scope 1 and Scope 2 data are available, 90 are assigned higher PDs and 82
lower PDs than with the use of sectoral data. The number of firms with higher and lower PDs rises to 227
and 205, respectively, when we employ the full sample of 432 firms for which individual data on either
Scope 1 or Scope 2 emission data are available, through the conversion of energy consumption into
emission levels.37 These results confirm the importance of individual emission data in capturing firmspecific risks that sectoral averages fail to reflect, particularly in high-emission industries where TR is
most pronounced.
Figure 14 illustrates the relation between PD estimates derived from granular and sectoral emission data
(‘granular PDs’ and ‘sectoral PDs’) under the stress-testing framework38. Although the median values are
comparable across the two datasets, the wider dispersion in granular PDs underscores the capacity of firmlevel data to reflect a broader array of risk exposures39. This variability reflects the improved detection of
vulnerabilities specific to individual firms, especially those in high-risk sectors40.
Namely, 227 firms showed higher PD estimates based on granular emissions compared to those derived from sectoral
averages, while 205 displayed lower PD values.
Transition risk-adjusted PDs are computed by monetising firm-specific emissions through a carbon bill, defined as the
product of excess emissions and a stressed carbon price. The resulting cost is propagated through the income statement
and balance sheet via accounting rules, and the updated financials are fed into the ICAS statistical model to estimate the
adjusted PD. See Cugliari et al. (2024) for methodological details.
The Kolmogorov-Smirnov test yields a significant p-value (2.28 x 10⁻⁶), confirming that PD distributions from detailed
and aggregated emissions data differ. Similarly, the Mann-Whitney U test p-value (9.67 x 10⁻⁷) supports the enhanced
discriminatory power of detailed emissions data. Quartile analysis further emphasises this disparity, with PD estimates
based on detailed data consistently showing higher values across all quartiles compared to those derived from aggregated
data.
The individual methodology’s ability to detect outliers is particularly evident in the upper quartile, where the data reveal
firms with significantly higher PD deltas under climate change stress scenarios. This finding helps demonstrating the
asymmetric nature of climate transition risks, where a subset of firms faces disproportionately high exposure. The sectoral
data, limited by their aggregate nature, fail to capture these nuances, thereby underestimating potential vulnerabilities.
Figure 14
Absolute difference between 1-year standard and risk-adjusted PDs
(percentages)
Source: authors’ TR-adjusted PDs calculations based on sectoral or individual emission data.
Note: the left portion of the figure shows, for each firm, the absolute difference between the one-year (standard) PD and the one-year riskadjusted PD when transition risk is proxied by sectoral averages (“Sectoral PDs”). The right portion shows the absolute difference between
the one-year (standard) PD and the one-year risk-adjusted PD when transition risk is calculated through granular emissions (“Granular
PDs”).
Table 4 and Table 5 show the difference in terms of ICAS ratings and CQS migrations produced by the
integration of transition risk in the credit risk assessment. Granular data result in a higher proportion of
downgrades (25 per cent) compared to sectoral estimates (14 per cent). This outcome reflects the
circumstance that granular data, by capturing firm-level differences, allow for the identification of
transition risks that are masked when using sectoral averages data. Stable ratings, referring to firms
for which the stressed PD under climate risk does not result in a rating migration, remain the vast majority
in both cases, accounting for 75 per cent in granular assessments and 86 per cent in sectoral ones.
Table 4
1-year change in credit ratings
(percentage values)
Method
Stable
Downgrade
Sectoral
Individual
Source: authors’ calculations based on sectoral or individual emission data.
Note: Table 4 illustrates the changes in credit ratings resulting from the climate stress test, reflecting how the ICAS baseline probability of
default (PD) was adjusted under climate risk scenarios. Table 5 presents the corresponding variations in credit quality steps (CQS), which
were also derived from the stress-induced shifts in PDs. The stress is inherently adverse, as the extra carbon bill (either imputed or reported
through the survey) resulting from the carbon tax negatively impacts financial indicators, leading to a deterioration in the PDs.
Table 5
1-year change in credit quality step
(percentage values)
Method
Stable
Downgrade
Sectoral
Individual
Source: authors’ calculations based on sectoral or individual emission data
In terms of CQS (table 5), not surprisingly the downgrades are less frequent than those observed for rating
changes, reflecting the less fine-grained nature of the CQS scale. However, individual assessments still
show a higher proportion of downgrades (15 per cent) compared to sectoral assessments (5 per cent).
These findings confirm that, by capturing firm-specific variations and extremes in transition risk, finegrained assessments allow for a more precise evaluation of credit risk, aligning with the goals of the Banca
d’Italia’s forward-looking risk management framework.
6.2 Physical risk
To evaluate the potential impact of the survey data, we adapt the ICAS methodology used to assess
physical risk exposure.
This approach starts from the baseline scores provided by an external provider, which estimate individual
firms’ exposure to floods and landslides. These scores41 are refined using additional information from the
ICAS survey. Specifically, the baseline score is adjusted upward (higher risk) when survey responses
indicate that recent physical events have caused significant damage – particularly when these events have
disrupted business operations in the short term. Conversely, the assessment is revised downward (lower
risk) when the survey responses indicate mitigation measures, such as insurance coverage, contingency
plans, or other resilience-enhancing actions. Hence, the final evaluation is an integrated view that
combines the starting score with firm-reported vulnerabilities and adaptive capacity into an ICAS physical
score.
This refinement is performed for the 511 firms in the survey that provide the necessary information. The
results show that the incorporation of additional information generally leads to a different assessment of
physical risk exposure. The resulting exposure distribution reduces the firms classified in the higher risk
classes and increases those classified in the negligible class, underlying an overall shift towards lower risk
classes (Table 6). This shows the value added of collecting data directly from individual firms, in order
to integrate the information available to commercial providers in assessing exposure to physical risk.
The external provider provides a discrete indicator of physical risk on a five-point scale (1 to 5), where higher values
indicate greater exposure. The score captures the riskiness of the areas where firms’ headquarters and operational units
are located, based on the likelihood and impact of various natural hazard events. The assessment is derived from geolocated data on company premises, linked to census tracts, and incorporates risk from acute climate events, chronic
climate phenomena, and non-climate physical hazards. The classification accounts for the type of facility exposed and
sector-specific vulnerabilities. Data sources include ISPRA, Copernicus, INGV, and other public datasets.
Table 6
Exposure to physical risk
(units; percentage values)
Exposure level
Baseline
Number of firms
Adjusted
Share
Number of firms
Share
Very high
Medium
Negligible
Total
100.0
100.0
Source: authors’ calculations based on ICAS Survey and external provider’s data
7. Conclusions
The ICAS survey carried out by Banca d’Italia in 2024 examines the climate-related risk management
practices of Italian NFCs within the context of the ICAS climate risk methodology. The survey addresses
some limitations of sectoral approximations, dictated by data availability, by collecting granular data on
emission strategies, risk management practices, and governance structures.
The survey identifies notable differences in climate risk management practices across sectors,
geographical areas, and credit quality classes. Governance structures related to climate issues vary
significantly across firms; deficiencies are more pronounced in the agriculture and services sectors, while
manufacturing firms generally exhibit more structured approaches. Nearly 43 per cent of firms have set
emission reduction targets over the next five years, yet no strong correlation emerges between these
commitments and firms’ credit quality scores. The measurement of emissions is limited, especially for
firms in sectors with medium and high emissions, and some firms quote the irrelevance of climate change
risks as a reason for not monitoring greenhouse gas emissions. 58 per cent of firms without climate
insurance view climate risks as negligible; the introduction of the compulsory insurance against damages
resulting from catastrophic events should mitigate this vulnerability. Firms with higher credit standing tend
to be more engaged in emission reduction. The results indicate that Italian firms have room for significant
advances in climate risk integration within their strategies, including through credible commitments to
long-term mitigation goals.
The use of the survey data as inputs for the ICAS model leads to improvements in both transition and
physical risk assessment, given that without those the assessment would be based on sectoral data. The
improvements are especially valuable for firms that do not provide non-financial disclosure. By
integrating granular emission data into the ICAS methodology, the analysis reveals significant
discrepancies between the PD estimates obtained with granular data and those that proxy them with
sectoral data.
Overall, the survey data provide an additional layer of analysis for the expert assessment performed by
the credit analysts which can more accurately assess the impact of physical and transition risks on the
creditworthiness of a firm with a new and direct estimation of firms’ vulnerability to climate-related risks,
as well as with a direct assessment of key governance and organizational variables. The findings of the
survey strengthen considerably the ability of ICAS to integrate sustainability considerations into its credit
risk assessment process by providing granular bottom-up information for climate risk analysis. This
enhancement allows to meet the Eurosystem common standards for the assessment of CCR that will be
implemented for the first time in 2025.
References
American Association for Public Opinion Research (AAPOR, 2023), Standard Definitions, Final
Dispositions of Case Codes and Outcome Rates for Surveys, 10th edition.
Angelini P. (2022a), Gli effetti dei cambiamenti climatici in Italia: strategie di adattamento e ruolo delle
imprese assicurative, Venezia, 28 ottobre 2022.
Angelini P. (2022b), The financial risks posed by climate change: information gaps and transition plans,
address by Paolo Angelini speech at Associazione nazionale per lo studio dei problemi del credito, Milan,
15 November 2022.
Angelini P. (2023), SMEs and the climate and environmental transition, speech at the Confindustria
Conference Finance and ESG Disclosure, Rome, 26 September 2023.
Auria L., Bingmer M., Caicedo Graciano C.M., Charavel C., Gavilá S., Iannamorelli A., Levy A.,
Maldonado A., Resch F., Rossi A.M. and Sauer S. (2021), Overview of central banks’ in-house credit
assessment systems in the euro area, Markets, Infrastructures, Payment Systems, N. 13, Banca d’Italia,
November.
Ballin M., Falorsi P.D., Falorsi S. and Pallara A. (2000), Il trattamento delle mancate risposte totali nelle
indagini Istat sulle famiglie e sulle imprese: soluzioni attuali e linee di ricerca, in Quintano C. (a cura di),
Scritti di Statistica economica 7, Quaderni di discussione, ISM-IUN, n. 19, Napoli.
Batteson C., Bevan A., Bunn P. and Taylor J. (2014). Quantifying catastrophic and climate-impacted
hazards. Bank of England Quarterly Bulletin, Q2 2014, pp. 219–226.
CCRA (2021), The Third UK Climate Change Risk Assessment Technical Report, London.
Cugliari M., Iannamorelli A. and Vassalli F. (2024), The dual face of carbon emissions: transition riskadjusted probability of default, Markets, Infrastructures, Payment Systems, N. 59, Banca d’Italia, July.
Di Virgilio S., Faiella I., Mistretta A. and Narizzano S. (2024), Assessing credit risk sensitivity to climate
and energy shocks: Towards a common minimum standards in line with the ECB climate agenda, Journal
of Policy Modeling, May.
ECB, (2022), Common minimum standards for incorporating climate change risks into in-house credit
assessment systems in the Eurosystem, Issue 6, Economic Bulletin.
Faiella I., Lavecchia L., Michelangeli V. and Mistretta A. (2022), A climate stress test on the financial
vulnerability of Italian households and firms, Journal of Policy Modeling, vol. 44(2).
Ferrando A., Groß J. and Rariga J., Climate change and euro area firms’ green investment and financing
‒ results from the SAFE in Economic Bulletin, Issue 6, ECB, 2023.
Frigo A. and Venturini A. (2024), La copertura assicurativa contro i rischi naturali: un’analisi
preliminare, Questioni di economia e finanza, n. 830, Banca d’Italia, Febbraio.
Gallo R., Guazzarotti G, Nigro V. e Cosconati M. (2022), Le coperture assicurative contro i rischi
operativi delle imprese italiane: alcune evidenze, Note di Stabilità finanziaria e vigilanza, n. 31, Banca
d’Italia, Ottobre.
Giovannelli F., Iannamorelli A., Levy A. and Orlandi M. (2020), The in-house credit assessment system
of Banca d’Italia, Occasional Papers, No. 586, Bank of Italy.
ISPRA (2023), Efficiency and decarbonization indicators in Italy and in the biggest European Countries.
Kunreuther H., Heal G., Allen M., Edenhofer O., Field C. B. and Yohe G. (2014), Integrated Risk and
Uncertainty Assessment of Climate Change, Climate Change 2014: Impacts, Adaptation, and
Vulnerability. Part A: Global and Sectoral Aspects. Contribution of Working Group II to the Fifth
Assessment Report of the Intergovernmental Panel on Climate Change (IPCC), pp. 151–194. Cambridge
University Press.
Lavecchia L., Appodia J., Cantatore P., Cappariello R., Di Virgilio S., Felettigh A., Giustini A., Guberti
V., Liberati D., Meucci G., Piermattei S., Schimperna F. e Specchia K. (2022), Dati e metodi per la
valutazione dei rischi climatici e ambientali in Italia, Questioni di economia e finanza, n. 732, Banca
d’Italia, Novembre.
MoF Platform on Sustainable Finance (2024), Sustainability Dialogue between SMES and Banks,
Ministry of Economy and Finance.
Monasterolo I. (2020), Climate Change and the Financial System, Annual Review of Resource
Economics.
Narizzano S., Orlandi M. and Scalia A. (2024), The Bank of Italy’s statistical model for the credit
assessment of non-financial firms, Markets, Infrastructures, Payment Systems, No. 53, Bank of Italy.
Rising J., Tedesco M., Piontek F. and Stainforth D.A. (2022), The Missing Risks of Climate Change,
Nature, October.
Smith T. W. (2023), A Revised Review of Methods to Estimate the Status of Cases with Unknown
Eligibility, Report of the Standard Definitions Committee for the American Association for Public
Opinion Research.
Appendix
Table a1
Rating scale
(percentage values)
Risk Class of
ECAI Scale
Minimum
Maximum
S&P / Fitch
Moody’s
0.000%
0.001%
0.001%
0.010%
0.010%
0.030%
0.030%
0.050%
0.050%
0.070%
0.070%
0.090%
0.090%
0.100%
0.100%
0.170%
BBB-H
0.170%
0.300%
0.300%
0.400%
BBB-L
0.400%
0.800%
0.800%
1.000%
CQS 4
1.000%
1.500%
CQS 5
1.500%
2.000%
2.000%
3.000%
3.000%
5.000%
B2/B3
B/B-L
5.000%
25.000%
CCC/C
Caa/C
CCC/C
25.000%
100.000%
Default
Default
Default
Eurosystem Credit
Quality Step
CQS 1&2
CQS 3
CQS 6
CQS 7
CQS 8
Table a2
Composition of the ICAS sub-sample
(unit; percentage values)
Number
Share
North West
North East
Center
South & Islands
Micro
Small
Medium
Large
1,386
Geographical Area
Revenues