Sunday, 30 November 2014

Effect of Employment Factors on Bondora Default Rates


I'm Brett, and in this article I'll present some more findings from an analysis of the Bondora Loan dataset.

Bondora is a P2P Lending Platform where lenders can potentially earn over 20% annual interest on their savings. That's nice in theory, but you're loaning your money out to people via unsecured loans, which can be risky!

So in this article I'll look at borrower employment factors, and whether they're a good indicator of likely loan default rates.

All of the factors I look at here are available in the Search Filter on the Bondora Market and Secondary Market screens.

Employment Status

See below for the chart on employment status:
Percentage default rates (Y) plotted against borrower's employment status (X)

The most noticeable thing here is that self-employed borrowers appear to have a much higher risk of defaulting on their loan obligations, and "entrepreneurs" have a much lower risk of default.

I'm not sure what the distinction Bondora makes between the self-employed and entrepreneurs. But this might go along with my previous analysis that suggested that borrowers borrowing for business purposes seemed to have a lower level of default.

Other than that, there seem roughly consistent default rates between employed, part-timers and retirees.

Occupation Area

The type of job a borrower has seems to have an influence the likelihood of them defaulting:
Percentage default rates (Y) plotted against borrower's occupation type (X)
I'm not sure I would read too much into this chart though. Industries come and go. For example, at the time I'm writing this we've witnessed the price of oil dropping from $110 to less than $70 in less than a year. That will likely lead to a lot of redundancies in the previously booming energy sector.
Other industries are notoriously cyclical, like IT in which I work.

Employment Duration with Current Employer
Percentage default rates (Y) plotted against borrower's time with their current employer (X)
As can be seen from the red trend line, there is a rough relationship between time with current employer and the chance of a default. However, without running the figures through a statistics package, I'm not sure how significant the differences are.

There is another category called trial period and there were no defaults amongst borrowers in this group. So these borrowers might be worth seeking out. However, there was quite a small sample size of just 58 loans for borrowers in this category, whereas the other categories had between 450 and 2700 loans.

Work Experience Years

There's a pretty good relationship between a borrower's number of years work experience and their likelihood of defaulting on their loans:
Percentage default rates (Y) plotted against borrower's years of total employment experience (X)
However, this factor is probably just a proxy for a borrower's age; my previous analysis showed that age has a big influence on the borrower's likelihood of defaulting on their loan obligations.

Current Job Title

Unfortunately I couldn't find this factor in the loan data spreadsheet, so there is no data available for this factor.

Data Generation

To generate this data, I used the following process:
  • I downloaded the loan Excel spreadsheet from Bondora.
  • I imported the Excel spreadsheet into SQL Server.
  • I wrote some custom SQL queries to analyse the data.
  • I exported the results sets from SQL Server back into Excel in order to turn them into charts.
  • I have assumed that the AD column equaling 1 indicates that a loan has defaulted. I have excluded loans that were applied for within the last 3 months or so. Finally, I've only included Estonian loans in all the queries except for the one relating to country.
If you want to have a go at analysing the data yourself, then this is the basic SQL query I used, in this case the query for the occupation_area factor:

    case occupation_area when 1 then 'Other'
    when 2 then 'Mining'
    when 3 then 'Processing'
    when 4 then 'Energy'
    when 5 then 'Utilities'
    when 6 then 'Construction'
    when 7 then 'Retail and wholesale'
    when 8 then 'Transport and warehousing'
    when 9 then 'Hospitality and catering'
    when 10 then 'Info and telecom'
    when 11 then 'Finance and insurance'
    when 12 then 'Real-estate'
    when 13 then 'Research'
    when 14 then 'Administrative'
    when 15 then 'Civil service & military'
    when 16 then 'Education'
    when 17 then 'Healthcare and social help'
    when 18 then 'Art and entertainment'
    when 19 then 'Agriculture, forestry and fishing'
    end 'Occupation Area',
    (Sum(AD) / Count(*) * 100) AS 'Percentage Defaulted',
    SUM(AD) as NumberInDefault,
    COUNT(*) as NumberOfLoans
From Loans
where country = 'EE' and creditdecision = 1
and occupation_area is not null and occupation_area > 0
and LoanApplicationStartedDate < '2014-10-27'
group by occupation_area
order by convert(int, occupation_area)

Summary and Conclusions

Once again there are some useful factors available that can help you to reduce your possible future default rates on Bondora.

Just bear in mind that economies are cyclical and industries come and go. So this can have a big influence on the ability of people to repay their loans, especially if they can't easily find employment in an alternative sector. Mining would have been booming when the price of gold hit $2000 and ounce, but at the time of writing it's back down to $1200, and could go a whole lot lower. 

Comments? Questions? Suggestions? Leave feedback below!

No comments:

Post a Comment