The ABC's of financial education: experimental evidence on attitudes, behavior, and cognitive biases

• Description of the programme
• The study
• Key findings
• Points to consider

Description of the programme

There is a growing consensus, amongst policy makers and NGOs, that financial education needs to be expanded, following the global economic crisis. However, evidence supporting the effectiveness of financial education is limited, in particular causal analysis is scarce and offers mixed results.

The World Bank developed the following field experiment to fill this gap, by capturing causal evidence of the effectiveness of financial education, and exploring the extent to which this is enhanced when combined with other interventions:

• Monetary incentives – participants took part in a ‘pay for performance’ knowledge test
• Behavioural nudges – participants set short-term, achievable goals and time-bound financial goals.
• Financial counselling - trained counsellors helped participants, in their own homes, to undertake a financial task such as preparing a budget or opening a bank account.

Each training session lasted 2-3 hours over five consecutive weeks. Participants were shown a video on the following topics: budgeting, savings, loans, insurance. After the video, the class (around 20) took part in a discussion led by a trained instructor. The videos used popular local actors and scripts were tested with local audiences.

The programme achieved an attendance rate of close to 100%. The authors do not comment on why this might be the case but they do note that all participants were already engaged with a local NGO, either in receipt of micro-finance or accessing another urban development programme.

The study

The study was conducted in Ahmedabad, an urban city in India, over four waves. The evaluators adopted a randomized evaluation approach, with a large sample of over 1300 participants. Two thirds of the sample were allocated to a Financial Education treatment group and one third was allocated to a control group. This group received a health education programme of the same intensity, so that both groups experienced similar levels of disruption to control for Hawthorne effects (whereby participants who know they are being studied are prone to modifying their behaviour).

There were also three additional treatment interventions delivered on a randomized basis, allocated in the following ways:

• Half of all participants in both the control and test groups were offered monetary incentives for correct answers on tests relating to either financial knowledge or health, dependent on which training sessions they attended.
• Half of all participants in the financial education treatment group were offered financial counselling, independent of whether they were also assigned to the goal-setting group
• Half of all participants in the financial education treatment group were allocated to the goal-setting group, independent of whether they were also offered financial counselling.

Within the treatment groups, there were groups who had received: financial education only; financial education with a monetary incentive; financial education with goal setting, and a very high intensity treatment which included all three components.

Respondents received Rs. 50 (US$1) to attend each session and transportation was free.

There were three data collection points: a baseline survey, a post-intervention knowledge survey, and a final survey which captured behavioural outcomes.

Key findings

• Financial Education alone was found to have positive effects on financial attitudes and awareness, but the addition of goal-setting and/or financial counselling was necessary to achieve behaviour change.
• Financial Education had strong positive effects on financial awareness (7% higher than the control group) and attitudes (8% higher than the control group), but no impact on financial numeracy.
• Combining financial education with goal-setting brought about ‘simple’ behavioural change such as: writing a budget, starting informal savings and avoiding borrowing for unforeseen expenses. Whereas the addition of counselling led to more ‘complex’ changes such as ongoing budgeting and holding a formal bank account.
• Financial incentives did not lead to any significant positive effects on financial attitudes and awareness.

• Budgeting: Financial Education had significant effects on understanding the benefits of budgeting (22% higher than control) and attempting to make a budget (28% higher than control), however the effect on ongoing budgeting was small (only 3% higher than control). The effect of financial education classes combined with goal setting was limited to raising awareness of budgeting and again, did not cause regular behaviour change. However, adding counselling to the former treatments caused a significant effect on sustained behaviour change. Those who received financial education with personal counselling were 38% more likely than the control group to have attempted a budget and 4% more likely to make a regular budget. This effect was even greater amongst those also in receipt of goal setting (43% and 5%).

• Saving: Financial education alone did not affect savings behaviour.Combing financial education with goal-setting increased the likelihood of informal saving (4.7% more likely than the control), counselling had no effect on informal savings but did significantly improve the likelihood of opening a savings account (13.8% more likely than the control group). The author states that counselling appears to enable participation in the formal financial sector.
• Borrowing: The effects of financial education alone on household borrowing were modest. When combined with goal-setting, participants were 2.5% less likely to borrow for unforeseen expenses. When all three treatments were combined, participants were 15% more likely to understand interest rates for their loans.
• Insurance: Impact on insurance was very low, but when all three treatments were combined, participants were 5.4% more likely to purchase life insurance.
• Providing classroom-based financial education alone generally yielded smaller results than when combined with higher-intensity, more personalised treatments, particularly in terms of long-term financial behaviour.

Points to consider

Methodological considerations

• This is a robust high quality randomised control study and as such offers strong causal evidence of the effectiveness of financial education and how this is enhanced when combined with other interventions.
• Attrition was exceptionally low at 6%, therefore we can be confident that findings are representative of the participant population.

Generalisability/ transferability:

• Study was undertaken in a developing country, participants were generally poor, with low very low levels of financial numeracy and education (only 4% had completed secondary school, for example) when compared to low income groups in developed countries.