Date post: | 12-Nov-2014 |
Category: |
Technology |
Upload: | venkatesan-g |
View: | 459 times |
Download: | 3 times |
Identification of Factors Influencing Financial Inclusion in Agriculture – An Application of Artificial Neural Networks
Vishnuprasad NagadevaraIndian Institute of Management [email protected]
Introduction
Financial development leads to economic growth
Financial development creates enabling conditions for growth
access to credit has a significant impact on agricultural growth
Introduction
Share of non-institutional sources of credit for cultivators had declined from 93% to 30% in recent years
Agricultural growth has slowed down, especially growth of food grains
Banks have been mainly focused on financing crop loans connected largely with food grains
Therefore, there is a reason to believe that financial exclusion may actually have increased in the rural areas over the last 10-15 years
Share of Credit (percentage)1951 1961 1971 1981 1991 2002
Cooperative Societies 3.3 2.6 22.0 29.8 30.0 30.2
Commercial Banks 0.9 0.6 2.4 28.8 35.2 26.3
Others 3.1 15.5 7.3 4.6 4.2 4.6
Total Institutional Sources
7.3 18.7 31.7 63.2 69.4 61.1
Money Lenders 69.7 49.2 36.1 16.1 17.5 26.8
Other Non-Institutional Sources
23.0 32.1 32.2 20.7 13.1 12.1
Non-Institutional Sources
92.7 81.3 68.3 36.8 30.6 38.9
Ratio of Bank Assets to GDP The ratio of bank assets to GDP is one of
the indicators of financial deepening. Indonesia : 101 per cent Korea : 98 per cent Philippines : 91 per cent Malaysia : 166 per cent UK : 311 per cent France : 147 per cent Germany : 313 per cent
India 80 per cent in 2005-06
Objectives of the study
Identify factors that influence the sources of credit for the agriculturists
To rank these factors in order of importance with respect to different sources of credit
Suggest appropriate policy measures to enhance financial inclusion
Methodology Factors that influence financial inclusion such as
gender, occupation, income groups, etc. are mostly either categorical or ordinal
Financial inclusion itself is a categorical variable. Chi-square is one of the Best techniques It is not amicable to determine the relative
importance Could not be used to prioritize the factors An alternate approach is needed One such technique is application of Artificial
Neural Networks
Artificial Neural Networks
Artificial Neural Networks (ANN) Process of Machine Learning Directed/Supervised Data Mining Applications in “Prediction”
Fraud Detection Customer Response Credit Rating
Neural Networks Mimic neurons of the human brain Links are the Processing Elements Learn from experience Good in detecting unknown
relationships PEs process data by summarizing and
transforming it through mathematical functions
Neural Networks PEs are interconnected and trained
and retrained repeatedly PEs are linked to inputs and outputs Training involves modifying the
weight or connection Uses “Learning Rules” to adjust
weights Training continues till desired
accuracy level is reached
Neural Network Model
Age
Region
Call Rate
Service
Income
Loyal
Hopper
Lost
Data The data is from the National Survey on Saving
Patterns of Indians. The sample covered both the rural and urban areas Various demographic characteristics such gender,
age, marital status, household size, education, profession, caste, asset ownership, media exposure
Information on coverage with respect to borrowings from different sources : financial institutions, money lenders, SHGs, relatives and friends etc.
The respondents belonging to the agricultural sector are selected
Sample Profile
Characteristic Frequency Percent
Age Group
Up to 30 1345 19.4
31-50 3782 54.4
51 & Above 1821 26.2
Total 6948 100
Gender
Male 6518 93.8
Female 430 6.2
Sample ProfileSocial Category
SC/ST 1800 25.9
Others 5148 74.1
Education Level
Up to Intermediate 2984 42.9
Graduation & Above 361 5.2
Marital Status
Currently Married 6189 89.1
Never Married 413 5.9
Widow/Widower 316 4.5
Divorced / Separated/Deserted 30 0.4
Sample Profile
Agricultural Landholding
Landless 709 10.2
Marginal 2829 40.7
Small 1626 23.4
Medium 1326 19.1
Large 458 6.6
Total 6948 100
Ownership of Occupied House
Yes 6617 95.2
No 331 4.8
Savings Instruments
Frequency Percent
Employee Provident Fund 5 0.1
Employees Pension Scheme 3 0
Government Pension Scheme 11 0.2
Government Provident Fund 12 0.2
Gratuity Scheme 11 0.2
Banking Products
Savings Account 3118 44.9
Fixed Deposits 332 4.8
Recurring Deposits 482 6.9
Banking Products
Savings Account 3118 44.9
Fixed Deposits 332 4.8
Recurring Deposits 482 6.9
Post Office Products
PPF 4 0.1
NSC 30 0.4
KVP 216 3.1
Insurance Products
Life Insurance (Endowment) 1013 14.6
Life Insurance (Non Endowment) 166 2.4
Personal Accident Insurance 32 0.5
Health Insurance 15 0.2
Non-Life General Insurance 47 0.7
Sources of Credit
Frequency Percent
No Credit 4890 70.38
Single Source 1528 21.99
Two Sources 372 5.35
Three Sources 76 1.09
More than 3 82 1.18
Total number of loans 2058 100.00
Important Sources of Credit
Source of Credit Frequency Percent Percent
Money Lender 947 13.6 46.02
Private Financial Institutions 86 1.2 4.18
Nationalized Banks 386 5.6 18.76
Cooperative Bank 337 4.9 16.38
Cooperative Society 303 4.4 14.72
Govt. 99 1.4 4.81
Relatives/Friends 623 9 30.27
SHG 116 1.7 5.64
Prediction Accuracies
Forecast Overall
Source Actual No Debt Debt
Money Lender
No Debt 86.07% 13.93%
84.71%Debt 17.00% 82.98%
Private Financial Institutions
No Debt 99.93% 0.07%
99.21%Debt 22.77% 77.02%
Nationalized Banks
No Debt 99.24% 0.76%
96.13%Debt 37.99% 61.90%
Prediction AccuraciesForecast Overall
Source Actual No Debt Debt
Cooperative Bank
No Debt 99.42% 0.58%
95.89%Debt 46.70% 53.18%
Cooperative Society
No Debt 99.50% 0.50%
98.34%Debt 14.26% 85.68%
Relatives/Friends
No Debt 98.35% 1.65%
93.80%Debt 37.03% 62.90%
SHG
No Debt 99.84% 0.16%
98.49%Debt 39.89% 59.83%
Variable CB CS ML NB PI RF SHG
Age Group 8 5 7 7 7 10 6Agricultural Landholding 6 3 1 2 9 4 5All Savings 9 Annual Expenditure 8 3 1 2 Annual Income 5 9 10 5 1 Annual Investable Surplus 6 1 Awareness of Alternative Investment Options
1 1 2 3 8 3 2
Variable CB CS ML NB PI RF SHG
Banking Products 10 7 7
Education Level 3 4 2
Exposure to Newspaper 9
Exposure to Radio 10 10
Exposure to TV 4 8
Insurance Products 2 8 4 3
Marital Status 4
Variable CB CS ML NB PI RF SHG
Media Exposure 2 9Owner of Other Real Estates 7 5 Post Office Instruments 9 Primary Savings Need 7 10 4 6 1Social Category 6 Language Proficiency [English] 8 3 Language Proficiency [Hindi] 9 6 5 8 10Language Proficiency [Local] 4 5 6
Table 8 helps in identifying eh unique factors There is a significant difference of the ranking Three factors - age group, agricultural land holding
and awareness of alternate investment options figure in all the sources
Social category (SC or ST ) only with respect to private financial institutions.
Annual expenditure is an important factor for credit from Cooperative societies.
Agricultural land holding is the most important factor with respect to money lenders,
Relative importance of these factors are likely to be different in different regions of the country.
The dataset used for the above analysis is an aggregate sample across the entire country
It is necessary to obtain different samples for different regions of the country so that the factors that are specific to different regions can be clearly identified
Questions? Suggestion? Comments?