Available via license: CC BY-NC-SA 4.0
Content may be subject to copyright.
https://doi.org/10.55343/CogSust.31
1 / 10
Environmental impact of mortgage bond purchases: presentation of a
possible estimation methodology
Balázs Lóránt
Magyar Nemzeti Bank
Budapest, Hungary
lorantb@mnb.hu
Gábor Fekete
Magyar Nemzeti Bank
Budapest, Hungary
feketeg@mnb.hu
Abstract
This article presents a methodological framework for estimating the environmental impact of mortgage bond
purchases. The model presented through the Mortgage bond purchase programme of the Magyar Nemzeti
Bank (MNB), the Central Bank of Hungary builds on the changing composition of the Hungarian housing
stock, and its main assumption is that, while maintaining the total floor area of the housing stock unchanged,
financing residential property modernises the housing stock as a result of tightening building energy
requirements, which reduces emissions. In our estimate, thanks to the MNB’s Mortgage bond purchase
programme, the Hungarian building stock could reduce its CO2 emissions by 13-41 thousand tonnes per year.
We have made several assumptions and simplifications in our calculations, and the results can only be
evaluated in this context.
Keywords
Mortgage bond, environmental impact, building stock, carbon emission
1. Introduction
Today, it is clear that climate change and sustainability are one of the greatest challenges facing humanity.
Every day there are news stories about temperature peaks, droughts or even floods caused by climate change.
Climate change affects all our lives in different ways and to different degrees. Accordingly, it is also essential
to look at how we are influencing climate change and the factors that cause it. Data and various calculations
and estimates play an essential role in mapping these. Without data, there are no calculations; without
calculations, there is no point in discussing effects.
In this article, we try to estimate the environmental impact of a previous asset purchase programme of
the Magyar Nemzeti Bank, the Mortgage bond purchase programme, using a model-based calculation. This
paper was inspired by the MNB’s Climate-related Financial Disclosure, the TCFD report (MNB, 2022a). We
started to address the issue in the context of this report. We believe that, despite the methodology and the
estimation difficulties, it can be a useful input for the analysis of other assets or asset purchase programmes,
and that is why we have written this article.
In the article, we first briefly review the, mostly technical, literature and the Mortgage bond purchase
programme itself. We then present the methodology and the model’s logic, assumptions and estimation
results. Finally, we summarise our writing.
2. Literature, background
Renewing and financing the renewal of housing stock is a goal for many countries worldwide.
Unsurprisingly, the built environment is a major contributor to emissions, accounting for almost half of the
annual global CO2 emissions. Of those total emissions, building operations are responsible for 27% annually.
10/10
In 2040 approximately two-thirds of the global building stock will be buildings that exist today, so without
improvements, it will continue to burden the environment (Architecture2030, 2022).
Renewing the building stock is also a major challenge in Hungary, where the housing stock is outdated
and has poor energy performance (Nagy, T. & Winkler, S., 2021). In this context, 40 per cent of the final
energy consumption in Hungary is related to buildings (MNB, 2022b). In their comprehensive study on the
Hungarian green bond market, Becsi et al. (2022) also highlight the high share of real estate in emissions and
the need for modernisation in Europe and Hungary.
Due to the specific nature of the study, the available literature on the environmental impact of securities
is very limited. Generally, we found that one direction of dual materiality, the financial materiality approach,
appears in the literature. By dual materiality, we mean the concept that not only does climate change have an
impact on companies (and other institutions), but they also have an impact on the climate, as shown in
Figure 1:
Figure 1. Double materiality
(Source: European Commission, 2019)
There is more literature on the former branch of dual materiality. Climate change's impact on
companies and the prices and risks of securities and other financial instruments has been the subject of several
studies, and the TCFD reports mentioned above typically follow this logic
1
. Kolozsi et al. (2022) conducted
a comprehensive study on this topic, analysing TCFD reports from central banks and examining the
measurement options and methodological challenges of climate risk regarding financial instruments.
1
It is important to highlight that the TCFD (Task Force on Climate-Related Financial Disclosures) provides recommendations, so although the
reports are similar in structure, they may differ in methodology and depth.
10/10
There is much less literature on the other direction of dual materiality, specifically on financial
instruments' environmental impact. This and the unavailability of data led us to try a new methodology in this
article. However, besides the lack of literature, it is also important to note that environmental impact reports
already exist, which, although not considered literature, contain much useful technical information on green
mortgage bonds. We also show examples of these for comparison when presenting our results.
3. Brief description of the MNB’s Mortgage bond purchase programme
In 2018, the Magyar Nemzeti Bank entered the mortgage bond market as a buyer under a programme,
and in 2020, to mitigate the negative economic impact of the coronavirus pandemic, it reintroduced its
Mortgage bond purchase programme. These two phases of the programme successfully supported the
objectives pursued: domestic mortgage bond issuance increased substantially, and the MNB helped banks to
obtain long-term, stable forint funding. Under the programmes, the MNB purchased fixed-rate mortgage
bonds denominated in forint from domestic mortgage banks and strongly emphasised participation conditions
enhancing transparency.
In the programme's first phase, the MNB made purchases in the primary and secondary markets for HUF
381 billion in nominal value throughout 2018 and HUF 307 billion in nominal value in the second phase,
from May 2020 to July 2021. In this article, we show how the environmental impact of mortgage bonds on
the central bank’s balance sheet can be estimated and the exposure of the MNB.
2
Mortgage bonds are securities issued against a specific regulatory background. Following Hungarian law,
at least 80 per cent of their collateral consists of so-called ordinary assets and the remainder of margin assets.
An ordinary asset is a principal amount outstanding from mortgage loans and interest due under the contract
3
.
These loans are backed by real estate as collateral, so the mortgage bond purchases indirectly finance the
construction, purchase, modernisation and extension of real estate. However, the link between mortgage bond
purchases and real estate is indirect, as the common collateral pool does not allow a clear link between the
actual transactions financed by the mortgage bonds and the contribution of MNB purchases to these
transactions. A further challenge is that, before the emergence of the MNB’s green initiatives, banks typically
did not collect data on the energy performance and environmental impact of the properties in connection with
the real estate transactions, so this data was not available for us. Overall, data scarcity and limited data
availability strongly determine the methodology that can be used to estimate the environmental impact of
mortgage bond purchases.
4. Description of the methodology (model)
The ultimate aim of the model is to estimate how much CO2 emission reductions could result from the
non-green phases of the MNB’s Mortgage bond purchase programme. We try to illustrate this by looking at
the change in the composition of the Hungarian housing stock, linking it to total mortgage lending and the
MNB’s purchases of mortgage bonds.
The model is based on the National Building Energy Performance Strategy published in 2015 (Ministry
of National Development, 2015). The document presents, among other things, the composition of the
Hungarian housing stock in 2013, broken down into 15 types of residential buildings, and also includes the
total floor area (m2) and specific energy consumption (kWh/m2/year). We previously explored other data
sources, but this was the only one we found sufficiently reliable and appropriate. We used the following table
in our model:
2
A new phase of the Programme was later launched, the Green Mortgage Bond Purchase Programme. This programme already introduced reporting
requirements, which would allow participating banks to provide their own, presumably more accurate, estimates of the environmental impact of
their mortgage bonds. However, the incentive effect of the Green Mortgage Bond Purchase Programme will come later than the period under review,
so this phase of the programme was not taken into account.
3
See more details: Act XXX of 1997 on Mortgage Loan Companies and Mortgage Bonds – https://net.jogtar.hu/jogszabaly?docid=99700030.tv
10/10
Building type
Total floor area
(m2)
Specific primary
energy
consumption
(kWh/m2a)
Calculated annual
energy consumption
(kWh)
1.
family houses below 80 m2 before 1945
15,918,873
551
8,771,299,023
2.
family houses above 80 m2 before 1945
29,610,378
408
12,081,034,224
3.
the family house below 80 m2 1946–1980
25,746,455
517
13,310,917,235
4.
the family house above 80 m2 1946–1980
83,997,263
405
34,018,891,515
5.
family houses 1981–1990
39,914,396
336
13,411,237,056
6.
family houses 1991–2000
23,667,465
227
5,372,514,555
7.
family or terraced houses (1 to 3 apartments) after
2001
24,466,147
173
4,232,643,431
8.
multi-apartment buildings (4 to 9 apartments)
before 2000
17,471,243
312
5,451,027,816
9.
multi-apartment buildings (4 to 9 apartments)
after 2001
2,929,898
125
366,237,250
10.
multi-apartment buildings (10 or more apartments)
before 1945
14,066,410
344
4,838,845,040
11.
multi-apartment buildings (10 or more apartments)
1946–2000
10,260,214
299
3,067,803,986
12.
multi-apartment buildings (10 or more apartments)
11,346,937
244
2,768,652,628
13.
multi-apartment buildings (10 or more apartments)
1946–1980
16,174,606
218
3,526,064,108
14.
multi-apartment buildings (10 or more apartments)
after 1981
9,877,417
200
1,975,483,400
15.
multi-apartment buildings (10 or more apartments)
after 2001
11,392,046
100
1,139,204,600
Total
336 839 748
339
114,331,855,867
Table 1. Building types with total floor area and energy consumption
(Source: Ministry of National Development, 2015)
In our model, we have assumed that mortgage-financed housing construction, new housing purchases
and renovations are likely to improve (reduce) the environmental burden of energy use in the building stock
through changes in the composition of the housing stock. It is important to emphasise here that we do not
estimate the environmental effect of construction in our model, but only the impacts of compositional change
due to the operation of the building, mainly related to energy use.
The “Bilan-Carbon” carbon footprint calculation method was used to calculate the initial
environmental burden, which is compatible with existing standards (ISO 14064) and the GHG Protocol
4
. To
this methodology, using the emission factors from the Clim’Foot database, so-called estimated average
emission factors (tonne/kWh) were determined for different building types. To determine this, we estimated
the average factors for each building type in terms of the proportion of energy used annually for lighting and
heating/cooling (20–80 per cent). Based on the emission factors thus determined, an aggregate CO2 emission
4
The Bilan Carbone method was developed by the French ABC and ADEME, and it is based on the Base Carbone database, a publicly available
emission factor database. The GHG Protocol and the ISO standard provide the framework for calculating the carbon footprint, while the specific
emission factors are provided by databases compiled by research institutes.
10/10
was calculated for this baseline period of 2013.
5
In 2017, Hungarian emission factors were developed for the
Bilan Carbone method under the LIFE Clim’Foot programme.
Aggregate CO2 emissions were also calculated using another method. As a tool for this, the “emission
factor” value of each building type – energy rating pair was taken from the European (including Hungarian)
residential property emission database of the Partnership for Carbon Accounting Financials. The results were
lower baseline CO2 emissions in 2013, resulting in lower CO2 emission savings associated with the assumed
modernisation of the building stock. Considering that the most reliable information in the baseline database
is the m2 data for the dwelling types, it is preferable to use emission factors (tonne/m2) from this database to
obtain results that best approximate reality. We could work with this methodology from February 2022, when
the Partnership for Carbon Accounting Financials database became available.
In the next step, we used the model to try and reproduce how this initial state might have changed
due to mortgage lending over eight years until the end of the period in 2021. For this purpose, we used the
Hungarian Central Statistical Office’s data table on mortgage lending for housing purposes
6
, which
distinguishes seven lending purposes: construction, new home purchase, modernisation/extension, purchase
of a second-hand home, swing loan, loan conversion and other purposes.
In line with the modernisation of the housing stock, the stock of construction, new home purchase
and modernisation/extension loans are fully considered in our model, as these are sure to finance more energy
efficient and, therefore, “greener” properties due to increasingly stringent building regulations and energy
requirements over the years.
In the case of loans for the purchase of a second-hand home, under the assumption that the buyer is
likely to move into or rent out the property after some renovation, and thus energy efficiency improvements
are also made, we also include some of these transactions (at rates of 5, 10, 15, 20 and 25 per cent) (see later).
Swing loans, loan conversion and other loans were excluded.
7
Based on the above, the number of loans available to make the composition of the building stock
more favourable from energy-efficiency and, therefore, environmental point of view. In addition, to estimate
the change in composition, we used the annual price per square metre statistics of new dwellings (HUF/m2)
published by the Hungarian Central Statistical Office for residential properties.
This was calculated annually and then summed over the period to determine how many square metres
of the more modern property could be built or modernised with a given amount of mortgage loans. A very
important feature and simplification of the model is that the total floor area of the building stock is kept
unchanged. As the floor area of the more modern buildings was added, the floor area of all types was reduced
proportionately, so the total square metres of the housing stock remained the same. Only the composition of
the stock changed, becoming more modern. This gives us an essentially fictitious housing stock composition
for the period under consideration, and we calculate an aggregate CO2 emission for this condition.
As a last step, the difference between the CO2 emissions in the current period (2021) and the baseline
period (2013) is used to obtain the annual CO2 emission reduction due to the modernisation of the housing
stock thanks to mortgage lending.
5
Databases containing carbon footprint calculation emission factors: Base Carbone - https://bilans-
ges.ademe.fr/en/accueil/contenu/index/page/bc_introduction/siGras/0, GHG Protocol - https://ghgprotocol.org/life-cycle-databases, EcoInvent
Database - https://ecoinvent.org/the-ecoinvent-database/
6
18.1.1.16. Housing loans - https://www.ksh.hu/stadat_files/lak/hu/lak0016.html
7
In the case of swing loans, no property modernisation can be expected, and the inclusion of loan conversions would result in duplication. Other
loans (e.g. general purpose mortgage loans) are generally not for housing modernisation purposes, as the terms are less favourable than for housing
loans.
10/10
A certain part of mortgage lending was financed by the MNB’s Mortgage bond purchase programme,
so in our model, based on a simple assumption, the MNB also contributes to the reduction of CO2 emissions
in the proportion of the amount of mortgage lending financed by its mortgage bond purchases and the amount
of all mortgage loans in HUF.
Figure 2. Presentation of the logic of the model
(Source: own edit)
5. Assumptions and estimation difficulties in the model
Regarding the methodology and the model, it is important to point out that several simplifying
assumptions have been made. One of the key simplifications is that our calculations assume that the total area
of the total housing stock does not change. This is a necessary simplification for the model, based on the
assumption that the number of occupied properties where actual energy use takes place does not increase in
the case of stagnating or decreasing population. In addition, the time elapsed between the disbursement of
the loan and the construction of the property may also be a factor of uncertainty, but we do not believe this
introduces a significant bias.
In addition, there are parameters to the assumption to which the model and its results are sensitive. As
mentioned above, our calculations considered the number of mortgage loans used to purchase second-hand
dwellings at a rate of between 5 per cent and 25 per cent for the improvement and modernisation of the
composition of the housing stock. Another important factor is the energy efficiency of the “modern” square
metres in the stock. We used values between 100 and 120 kWh/m2/year in our calculations based on
information from the Hungarian Central Statistical Office.
In addition, we explore several possibilities for the distortion caused by the collateral behind mortgage
bonds. As described earlier, the underlying collateral for mortgage bonds must be ordinary assets (mortgage
loans) of at least 80 per cent, but the proportion of ordinary assets is typically higher (varying by mortgage
bond). The full possible range of 80 to 100 per cent was examined in this case. Furthermore, within ordinary
10/10
assets, loans are secured by residential and commercial real estate. Nevertheless, the collateral for loans is
typically residential property, in 98–99 per cent, so we do not include this in our analysis. Finally, it is very
important to point out that the theoretical CO2 emissions of the housing stock presented in the above model
may differ significantly from the actual emissions because the building stock includes vacant properties and
properties that are only used temporarily (holiday homes). In our model, therefore, we do not attempt to
calculate accurate CO2 emissions from actual energy consumption data based on the baseline (2013) or the
current period (2021) composition of the building stock but to estimate the environmental burden reduction
associated with the modernisation of the housing stock based on mortgage loan disbursement data for the
period.
6. Presentation of the results
Based on our calculations, annual emission avoidance of between 13 and 41 thousand tonnes is estimated
to result from the MNB’s Mortgage bond purchase programme by modernising the housing stock. It can be
seen that there is considerable uncertainty in the estimate, covering a relatively wide range, due to the
assumptions and different scenarios tested.
By comparison, the actual CO2 emissions from the cooling/heating of households in Hungary averaged
around 15 million tonnes per year in the period under review. In comparison, the average annual CO2
emissions from the operational activities of the Magyar Nemzeti Bank were around 6 thousand tonnes. Thus,
according to the model, the MNB’s Mortgage bond purchase programme can save Hungary approximately
2–7 times the MNB’s annual carbon footprint on an annual basis due to the effect of the housing stock
modernisation.
We want to organise and interpret the results. The best way to do this is to compare with impact reports.
The impact reports show the environmental impact of the resources invested in green projects. The impacts
are compared to an alternative scenario (in this aspect, they are similar to the model), which aims to show the
positive environmental change that can be achieved by implementing the green project. In the case of
mortgage bonds, this typically means the emissions saved through making the property more energy efficient.
Our results are compared with the impact reports of green mortgage bonds previously issued in Western
Europe.
8
These are the basis for the comparison: each tonne of CO2 emission avoided is calculated per year
per million euros
9
. This is illustrated in the figure below, where light blue represents our estimated range of
the MNB’s contribution to emission avoidance:
8
The green mortgage bonds of these financial institutions have similar characteristics to Hungarian mortgage bonds (except for their green
character), which is why they were chosen as a basis for comparison and because impact reports are available for these few institutions.
9
Using ECB exchange rate data: https://www.ecb.europa.eu/stats/policy_and_exchange_rates/euro_reference_exchange_rates/html/index.en.html
10/10
Figure 3: CO2 emissions avoided per 1 million euro in tonnes per year for certain green bonds and the MNB’s programme*
Source: MünchenerHyp
10
, DNB
11
, Sparebanken Sør
12
, Berlin Hyp AG
13
, MNB
*Note: The value reported regarding the MNB’s programme is calculated using a different, specific estimation methodology and several uncertainty
factors, so its comparison with the presented green mortgage bonds should be treated cautiously.
It can be seen that the MNB programme achieves a relatively high overall reduction in emissions per
million euros. This may be partly because the Hungarian real estate stock is starting from a more outdated
situation. Moreover, Hungarian real estate is relatively cheaper. Thus the amount of money invested may
result in a larger reduction in emissions. It is also important to underline that different institutions may have
different loan targets and property types behind their mortgage loans, with different criteria, which also affect
the presented results. Finally, we would like to draw attention again to the model's uncertainties and
assumptions. The comparison is an attempt to estimate the scale of the results of the MNB programme.
7. Summary
In our article, we attempted to estimate the environmental impact of the Magyar Nemzeti Bank’s
Mortgage bond purchase programme. After a brief introduction, we have attempted to review the relatively
limited relevant literature available on the environmental impact of securities. We then presented the MNB
Mortgage bond purchase programme, which was the subject of our analysis, and aimed to estimate its
environmental impact. Our methodology was based on the changes in the composition of the Hungarian
housing stock, and the main assumption was that, while maintaining the floor area of the housing stock
unchanged, the financing of housing would have a modernisation effect due to the tightening building energy
requirements and the renewal of the housing stock, which will lead to a reduction in emissions. In the
methodological section, we presented our sources alongside the logic of our model. The assumptions and
simplifications used in our calculations were then explained. These factors can significantly impact the
results, so the estimations can only be evaluated under these assumptions and simplifications. Thanks to the
MNB’s Mortgage bond purchase programme, we estimate that the Hungarian building stock can reduce its
CO2 emissions by 13-41 thousand tonnes per year.
10
https://www.muenchenerhyp.de/sites/default/files/downloads/2021-05/mhyp_Impact_Reporting_2020_en_02.pdf
11
https://www.ir.dnb.no/sites/default/files/3Q19%20Green%20Covered%20Bond%20Impact%20Report.pdf
12
https://www.sor.no/globalassets/organisasjon/2020-green-impact-report.pdf
13
https://www.berlinhyp.de/en/investors/green-bonds
10/10
References
Architecture2030. (n.d.). Why the build environment? URL: https://architecture2030.org/why-the-building-sector/ (Downloaded: 22 August 2022)
Becsi, A., Varga, M., Loga, M., & Kolozsi, P. P. (2022). First steps – the nascent green bond ecosystem in Hungary. Cognitive Sustainability, 1(1).
DOI: https://doi.org/j362
European Commission. (2019). New guidelines on reporting climate-related information.
URL: https://ec.europa.eu/info/sites/default/files/business_economy_euro/company_reporting_and_auditing/documents/190618-
climate-related-information-reporting-guidelines-overview_en.pdf 10 August 2022.
Kolozsi, P. P., Ladányi, S. & Straubinger, A. (2022). Measuring the Climate Risk Exposure of Financial Assets – Methodological Challenges and
Central Bank Practices. Financial and Economic Review, Vol. 21 Issue 1, March 2022, 113–140. DOI: https://doi.org/j363
Ministry of National Development. (2015). National Building Energy Performance Strategy.
URL: https://ec.europa.eu/energy/sites/ener/files/documents/2014_article4_hungary_en%20translation.pdf
MNB. (2022a). The Magyar Nemzeti Bank’s climate-related financial disclosure. URL: https://www.mnb.hu/letoltes/tcfd-jelente-s-2022-en.pdf
(Downloaded: 12 June 2022)
MNB. (2022b). Green Finance Report. URL: https://www.mnb.hu/letoltes/20220718-green-finance-report-2022-3.pdf (Downloaded: 01 August
2022)
Nagy, T. & Winkler, S. (2021). Előremutató követelményekkel érhetünk csak el igazi eredményeket a zöldítés terén (Forward-looking requirements
are the only way to achieve real results in greening). URL: https://www.mnb.hu/letoltes/nagy-tamas-winkler-sandor-zold-otthon-
program-bb-kriterium.pdf (Downloaded: 22 August 2022)