Version classiqueVersion mobile

Climate-Related Financial Risks for Kenyan Banks

 | 
Reuben Muhindi Wambui

3. Data and stylised facts

Texte intégral

3.1 Bank data

1Kenyan banks’ financial data are obtained from three sources: CBK Banking Annual Supervision reports and respective banks’ annual financial reports and investor presentations. My analysis uses 2018 lending data reported by banks, or 2017 data where 2018 data on sectoral loan composition is missing. Relative to the economy, the banking sub-sector assets accounted for 49.51% of Kenya’s nominal GDP.

2The CBK report provides the gross loans amounts for each bank and the industry gross loans across 11 lending sectors in the economy (Table 1). Overall, the industry value of gross loans in 2018 was KES 2,483.5 billion (Central Bank of Kenya, 2018), with the highest lending being to personal loans (26.6%) and the lowest to mining and quarrying (0.5%). There are four sectors with at least 5% of industry gross loans: trade (19.1%), real estate (15.2%), manufacturing (13.0%), and transport and communication (6.6%).

Table 1: Banking industry gross loans by sector, 2018

Lending Sector

Industry

Share

Personal

26.6%

Trade

19.1%

Real estate

15.1%

Manufacturing

13.0%

Transport and communication

6.6%

Energy and water

4.4%

Building and construction

4.1%

Financial services

3.9%

Agriculture

3.6%

Tourism, restaurants, and hotels

2.9%

Mining and quarrying

0.5%

Total

100%

3The analysis focuses on the banks’ lending portfolios because the breakdown of other banking assets by sector categories that can be directly mapped to GHG emissions is not available. Nonetheless, Kenya’s banking industry assets are mostly loans, so this analysis is appropriate. In 2018, loans and advances accounted for 52% of the banking industry balance sheet, followed by government securities (27%). Investments and other assets are a small 9% and the residual is made up of cash, placements, and balances at the CBK (Central Bank of Kenya, 2018).

Sample of banks

4Kenya had 40 registered banks in 2018, excluding two banks in receivership and one under statutory management. Of these, 13 are listed on the Nairobi Securities Exchange (NSE), and 12 of these are local banks and one is a cross-listing of Rwanda’s Bank of Kigali. The analysis uses a sample of 12 banks; this includes 11 listed Kenyan banks (all listed banks except HF Group) and Family Bank which, although not listed, accounted for 2% of industry gross loans. HF Group is omitted from the sample since it is primarily a property and mortgage lending bank and thus has no comparable sectoral breakdown in its reporting. The sample of banks used accounts for 82% of gross loans and I deem this satisfactorily representative of the Kenyan banking industry (Table 2).

Table 2: Market shares of industry gross loans, 2018

Bank

Market share

1 KCB

0.17

2 Coop

0.10

3 Equity

0.09

4 Barclays*

0.08

5 CFC Stanbic

0.06

6 DTB

0.06

7 I&M

0.06

8 SCB

0.05

9 CBA**

0.05

10 NIC**

0.05

11 NBK***

0.03

12 Family

0.02

Other banks

0.18

Total

1.00

*Barclays has since rebranded to Absa Kenya

**CBA and NIC have since merged to form NCBA Bank

***NBK has since been acquired by KCB Bank

Sectoral loan composition

5Although the CBK reports provide gross loans amounts for each sector at industry level and gross loans amounts for each bank, they do not provide the sectoral breakdown at bank level. Bank-level sectoral loan composition is necessary for the analysis of climate risk exposure by sector across banks. For this, I use sectoral lending breakdown as reported in individual banks’ annual financial reports or investor presentations. Some banks report the actual gross loans advanced to each sector in Kenyan shillings while some (e.g., Equity Bank) report the percentage of lending shares across the sectors from which actual gross loan amounts can be estimated.

6However, a discrepancy in reporting makes the gross loans amounts not immediately comparable across banks: banks do not report gross loans across similar sector categories, hence there is no standardised reporting of sectoral loan composition. For example, while Coop Bank reports lending across 10 sectors (the same as the 11 sectors in the CBK report but without mining and quarrying loans), Barclays Bank (now Absa Kenya) reports gross loans across 6 sectors: manufacturing; wholesale and retail trade; transport and communication; agriculture; private individuals; and others. The difference in sectoral reporting of gross loans across banks requires standardising the reporting across banks.

7In standardising the sectoral reporting, I make a number of adjustments:

“Manufacturing, energy, and water” sector: CBK publishes distinct gross loans advanced to the manufacturing sector and to the energy and water sector. Since many banks aggregate loans to these sectors, I aggregate bank-level loans to either of these sectors to one sector category.

“Real estate, building, and construction” sector: CBK publishes distinct gross loans advanced to the real estate sector and to the building and construction sector. Since many banks aggregate loans to these sectors, I aggregate bank-level loans to either of these sectors to one sector category.

“Personal loans” sector: Where a bank reports gross loans advanced to “others” but has no distinct personal/consumer loans, these loans advanced to “others” are treated as personal loans. This applies to CFC Stanbic, I&M Bank, NIC Bank, and Family Bank. I decided on this treatment after consulting Kenyan bankers working in credit departments; these loans are mostly personal loans to their customers whose use is not reported like loans to any of the specified sectors. All other banks have distinct personal loan categories. However, these are not similar to loans to “other services”.

8After standardisation of sectoral reporting across banks, there are nine distinct sector categories for each bank. Each of these sectors can be mapped directly to the CBK reporting and the output of the standardisation accounts for 96% of the gross loans in the sample. The remaining 4% belongs to loans to “other services” which could not be mapped directly to any of the 11 distinct sectors as reported by the CBK. Table 3 shows the standardised sectoral loan composition. Column A is the composition at industry level while Column B is that of the sample.

Table 3: Standardised reporting of sectoral loan composition, 2018

A

B

Industry

(CBK Report)

Sample (Banks’ Reports)

Personal

26.6%

29.8%

Agriculture

3.7%

4.2%

Mining and quarrying

0.4%

0.1%

Manufacturing, energy, and water

17.5%

14.7%

Real estate, building, and construction

19.3%

18.0%

Wholesale and retail trade

19.1%

15.2%

Tourism, restaurants, and hotels

2.9%

1.9%

Transport and communication

6.6%

7.4%

Financial services

3.9%

4.3%

Other services

4%

Total

100%

100%

9The two columns are consistent and the small discrepancies are explained by 1) the fact that the sample uses 12 banks (accounting for 82% of industry gross loans) and 2) the 4% residual of loans to “other services” within the sample. I proceed to use these sector categories for the analysis and Table 4 summarises the sectoral loan composition employed (extracted from column B of Table 3).

Table 4: Sectoral loan composition of sampled banks, 2018

Lending sector

Sector Share

Personal

29.8%

Agriculture

4.2%

Mining and quarrying

0.1%

Manufacturing, energy, and water

14.7%

Real estate, building, and construction

18.0%

Wholesale and retail trade

15.2%

Tourism, restaurants, and hotels

1.9%

Transport and communication

7.4%

Financial services

4.3%

Other services

4.5%

Total

100%

3.2 Emissions data

10The analysis uses the 2015 GHG emissions data as reported in the NDC Sector Analysis Report (Government of Kenya, 2017). Total national GHG emissions in 2015 were 80 metric tonnes of carbon dioxide equivalent (MtCO2e) as shown in Table 5. There is no more recent data and the emissions data for the subsequent years are forecast estimates. The NDC target is abating GHG emissions by 30% relative to the 2030 BAU scenario. The BAU forecast for 2030 is 143 MtCO2e.

Table 5: Kenya GHG emissions by sector, 2000–2030

Baseline Emissions (MtCO2e)

Sector

2000

2005

2010

2015

2020e

2025e

2030e

Agriculture

23

26

30

32

34

36

39

Electricity generation

1

1

1

1

12

24

42

LULUCF*

21

18

21

26

25

23

22

Transportation

4

4

7

9

12

16

21

Energy demand

5

5

6

7

8

9

10

Industrial processes

1

1

2

3

4

5

6

Waste

1

2

2

2

3

3

4

Total

55

57

70

80

98

115

143

*Land use, land-use change, and forestry

Note: Where the totals do not add up, this is due to rounding errors

11Table 6 is an extract of the sectoral contribution to the national emissions in 2015.

Table 6: Kenya GHG emissions by sector, 2015 (baseline)

Emission Sector

Emissions (MtCO2e)

Sector %

Agriculture

32

40%

Electricity generation

1

1%

LULUCF

26

33%

Transportation

9

11%

Energy demand

7

9%

Industrial processes

3

4%

Waste

2

3%

Total

80

100%

12Agriculture is the biggest contributor to Kenya’s GHG emissions (40%), followed by the land use, land-use change, and forestry (LULUCF) sector (32%), largely as a result of deforestation. Transport is the third largest contributor (11%).

3.3 Bank-level sectoral measures – Wij and Mij

13I define two measures of interest in each bank’s lending portfolio: 1) the share of each bank’s loan portfolio to each sector, Wij, and 2) the market share of each bank in each lending sector, Mij. These two measures are adopted from Monasterolo et al. (2017) who defined sectoral measures in constructing "GHG exposure" and "GHG holding" indices for portfolios of equity holdings and loans in the euro area.

The share of each bank’s loan portfolio allocated to each sector, Mij

14Let GLij denote the gross loans by each bank i to each economic sector j. Also, let the share of emissions attributable to each sector be denoted by Sj.

15Wij, being the share of each bank’s loan portfolio to each sector can be calculated as:

Image 100002010000013B0000006E8962428B972BF782.png (1)

16where the index k runs over the set of all sectors and where ΣWij = 1 for each bank.

17For example, if Bank A lends to only two sectors, 40% to sector 1 and 60% to sector 2, then WA1=0.4 and WA2=0.6.

18From this preliminary analysis, Table 7 below reports the values of Wij for the banks across the sectors. The highest financial exposures are seen mostly in four sectors: personal loans; real estate, building, and construction; manufacturing, energy, and water; and wholesale and retail trade. This is seen across most of the banks including the top three banks in market share of gross loans (KCB, Coop, and Equity). The lowest financial exposure is seen in the mining and quarrying sector, with only I&M bank reporting loans to this sector within the sample.

Table 7: Share of each bank’s loan portfolio to each sector, Wij

Image 10000201000006B00000049014F3C6435BD07EF6.png

*Barclays has since rebranded to Absa Kenya

**CBA and NIC have since merged to form NCBA Bank

***NBK has since been acquired by KCB Bank

NBK did not report any personal loans category in its sectoral breakdown but these are likely included in “other services” (reported as Business Services in the financials) that account for 42% of the bank’s loans

The market share of each bank in each lending sector, Mij

19Again, GLij denotes the gross loans by each bank to each economic sector, Sj is the share of emissions attributable to each sector, and Wij is the share of each bank’s loan portfolio to each sector.

20Mij, being the market share of each bank in each lending sector can be calculated as:

Image 10000201000001410000006EC549D6E053AC02CE.png (2)

21where the index k runs over the set of all banks and where ΣMij = 1 for each sector.

22For example, if Banks A and B are the only banks that lend to sector 2, accounting for 45% and 55% of loans to that sector respectively, then MA2 = 0.45 and MB2 = 0.55.

23From the preliminary data analysis, Table 8 reports the values of Mij for the banks across the sectors. The market shares in each sector do not add up to 1 since some market share in each sector is held by non-sampled banks. As expected, large banks have dominant market shares in the different sectors with KCB Bank having the largest market share in three sectors: personal; manufacturing, energy, and water; and real estate, building, and construction.

Table 8: Market share of each bank’s loan portfolio in each sector, Mij

Image 100002010000065E000004623713B1D6097DD5C6.png

*Barclays has since rebranded to Absa Kenya

**CBA and NIC have since merged to form NCBA Bank

***NBK has since been acquired by KCB Bank

NBK did not report any personal loans category in its sectoral breakdown but these are likely included in “other services” (reported as Business Services in the financials)

3.4 Mapping of emissions to bank lending sectors

24Having obtained bank-level measures, the next step is to map the bank lending sectors to the GHG emissions reported by the Ministry of Environment and Natural Resources. This is necessary since the emissions reporting sectors are not similar to the lending sectors and there is no national reporting of emissions by categories equivalent to the lending sectors. The mapping of GHG emissions to bank lending sectors is a key contribution of this study to the literature on climate financial risk in Kenya and other emerging economies.

25For the mapping of GHG emissions to bank lending sectors, I considered two treatments:

  1. One-to-one sectoral mapping: This is for lending sectors than can be directly mapped one-to-one to any of the emission reporting sectors. Four sectors qualify: agriculture; transport; electricity generation; and industrial processes. Collectively, they account for 45 MtCO2e (56% of emissions).

  2. GDP-weighted emission mapping: For the remaining 44% of emissions, I allocated them to the other lending sectors based on each sector’s contribution to GDP; i.e. allocation of emissions to each sector is based on each of the sector’s percentage contribution to national output. Data on sectoral contribution to GDP is obtained from the CBK reports and corroborated with data from the Kenya National Bureau of Statistics (KNBS). Historically, emissions have been strongly correlated with economic activity. Cohen et al. (2018) found that, while key developed economies show signs of the decoupling of emissions and output trends, in emerging economies there is still a strong upward trend in emissions that is matched by an upward output trend.

26Table 9 shows the mapping of the lending sectors to the six emission categories.

Table 9: Mapping of GHG emissions to bank lending sectors

GHG Emission Sector

Bank Lending Sector

Emissions %

- Agriculture

- Agriculture

40%

- Transportation

- Transportation and communication

11%

(2 sectors combined)

Electricity generation

Industrial processes

- Manufacturing, energy, and water

5%

(3 sectors combined)

LULUCF*

Energy demand

Water

Combined emissions split across 7 sectors by GDP contribution weighting:

Personal

Mining and quarrying

Real estate, building, and construction

Wholesale and retail trade

Tourism, restaurants, and hotels

Financial services

Other services

44%

Total

100%

*Land use, land-use change, and forestry

27It is worth noting that energy demand as an emission sector relates to national energy demand at all levels (household, industrial, and commercial) and thus is not be erroneously directly mapped to the manufacturing, energy, and water sector.

28Table 10 provides emissions weights of each banking sector based on the mapping.

Table 10: GHG emission contributions by lending sector, 2018

Sector

Emissions %

Personal

6%

Agriculture*

40%

Mining and quarrying

1%

Manufacturing, energy, and water*

5%

Real Estate, building, and construction

12%

Wholesale and retail trade

7%

Tourism, restaurants, and hotels

1%

Transport and communication*

11%

Financial services

6%

Other economic activities

11%

Total

100%

*These sectors were mapped directly to their respective emission sectors. The other sectors’ emissions contributions were determined from GDP-weighting as seen in Table 9

29“Other economic activities” in the emissions mapping refers to GDP-weighted emissions from economic activities not directly linked to any of the distinct bank lending sectors. These include the following economic activities captured in the GDP reporting by KNBS: public administration; professional, administration, and social services; education; health; and other services. Most of these are public services and it is reasonable that there are national emissions from public and social services not linked to the sectors under CBK reporting.

30Table 11 below replicates GHG emissions by sector with the total national emissions rescaled across all bank lending sectors without the “other economic activities” category. The redistribution keeps the relative weights across all the bank lending sectors and thus does not distort the analysis.

Table 11: GHG emission contributions by sector (rescaled emissions=100%)

Sector

Emissions %

Personal

7%

Agriculture*

45%

Mining and quarrying

1%

Manufacturing, energy, and water*

6%

Real estate, building, and construction

13%

Wholesale and retail trade

8%

Tourism, restaurants, and hotels

1%

Transport and communication*

13%

Financial services

6%

Total

100%

*These sectors were mapped directly to their respective emission sectors. The other sectors’ emissions contributions were determined from GDP-weighting as seen in Table 9.

CC-BY-NC-ND-4.0

Le texte seul est utilisable sous licence CC BY-NC-ND 4.0. Les autres éléments (illustrations, fichiers annexes importés) sont « Tous droits réservés », sauf mention contraire.

Acheter

Rechercher dans OpenEdition Search

Vous allez être redirigé vers OpenEdition Search