Two categories of regular readers are marketers who use research and people who conduct the research marketers use. Both groups often ask how to tell the difference between good information and bad. Because most research is privately funded for competitive purposes (proprietary), details cannot be shared. But published articles are fair game. Recently I read an article published by Priceonomics, whose tagline is “Turn your company data into content marketing that performs.” The study that provided the data was conducted by LendUp, who says they offer a better alternative to payday loans. The report is titled Why Do People Get Payday Loans? Before we take a closer look at the study, some background will help us establish a baseline from which we can compare the results.

What are payday loans?

They are unsecured, no-credit-check “signature” loans where lenders loan money at high interest rates to borrowers with low incomes and poor credit or none at all. Borrowers are expected to repay loans from their next paycheck. Their ability to repay loans while meeting their other financial obligations is generally not considered by the lender.

Who borrows from payday lenders?

Three out of four borrowers have full-time jobs, so it’s mostly people with annual incomes between $20,000 and $30,000 and who live from paycheck to paycheck. Most borrowers’. Forbes says 58% of payday loan borrowers have difficulty meeting basic monthly expenses like food, rent and utility bills.

Three-fourths of all payday loans are to borrowers who have taken out 11 or more loans in a year

Most payday loans are taken out within two weeks of repaying a previous one. In the United States, 12 million people use payday loans to borrow nearly $50 billion a year.

The Federal Deposit Insurance Corporation says 14 million U.S. adults are unbanked and 49 million adults are underbanked 

Unbanked means a person has no checking or savings account with any bank. Underbanked means a person has an account, but looks outside the banking system for loans. Generally speaking, payday loan applicants cannot get loans from banks and credit cards because they have no accounts with either.

There are two sides to the payday loan debate 

Advocates say payday loans:

• Are the only type of loan available to most low-income Americans.
• Provide “democratization of credit” by loaning to millions of poor people when other financial institutions will not. 
• Keep borrowers from having to use illegal sources. 
• Improve household welfare.
• Provide aid in disaster areas.
• Have lower profit margins than credit unions and banks.

Critics say:

• The industry takes advantage of those who are poor and/or in financial distress because borrowers aren’t always aware of the high interest rates they’re paying. 
• Pew Charitable Trusts says real interest rates range from a low of 196% in Minnesota to a high of 574% in Wisconsin and Mississippi.
• According to the Center for Responsible Lending, the average annual interest rate works out to about 400%.
• In 2013, the Consumer Financial Protection Bureau (CFPB) defined practices used by payday lenders as being unfair and abusive. CFPB director Richard Cordray says “Fees are exorbitant and amount to predatory lending. Consumers are being set up to fail with loan payments they are unable to repay.”
• A Federal Reserve Bank study says access to payday credit encourages borrowers to over-consume and spend less on such essentials as food and rent.

Take a look at the graphic below. What do you see?

When I look at it, I see many things that concern me, but one really stands out

What catches my eye right away is nearly half the reasons given in this study are reported only as “Other.” There are two likely explanations for combining so many unidentified variables into a miscellaneous category. 

1. Sometimes there are many responses that are only mentioned by a few people, and each of the percentages is quite small. Because we see 2.3% (Repay Another Loan) is deemed worthy of reporting, the implication is that each of the variables in “Other” are smaller than that.
2. The other is the design of the questionnaire did not allow for any categories other than the six we see reported. This is a very typical problem with surveys.

If the first is so, then we should expect to be told about some of the many responses that got put in such a large “Other” bucket, but we are shown none.

We also should expect to see a note on the slide that says something like “No individual responses were greater than 2%.” If the problem is a bad questionnaire, it means we can never know what things are hiding behind the “Other” door. In this instance, where a writer was handed data from a study done by someone else, there is a third alternative – only some of the data were passed along. When nearly half of survey responses are not identified, either survey designers have irretrievably buried meaningful data or writers are holding back information.

The rule of thumb all good analysts use is to collect detailed data and combine it into categories later

The problem with too many surveys is they determine the categories in advance because it is faster and cheaper to count numbers from clicks than it is to record things in people’s own words and convert them to statistics later. The written text to this report adds nothing that we didn’t already know from looking at the chart because it only tells us is what we already saw in the graphic. This style of reporting is merely descriptive, involves very little actual analysis and is the equivalent of reading slides to a roomful of people. You can judge for yourself by clicking here to read the Priceonomics report. Please note that they do not show us a methods section or the sample size.

The three other tables included in this article show variance by three demographic categories: Income, Age, and State

Let’s first look at a table I built myself so I could have a broad overview before looking at the details.

Do you see what I see when you look at these simple cross-tabulations? 

• The range of responses for variance by income is only 0% to 4% – surprisingly small, wouldn’t you say? 
• Age shows a spread of 17 percentage points (there may be something interesting here). 
• State (of residence) shows a ten-point spread. Note survey takers are from only 12 states, not all 50, indicating these are areas of particular interest to sponsors or a really sloppy piece of research – again, that information is kept from us.

The next table is titled Income and Reason for Getting a Payday Loan

What do you see when you take a closer look at it?

Knowing payday loan borrowers are people with low incomes, the report’s income “buckets” are quite surprising, aren’t they? We know nearly 80% of borrowers earn between $20,000 and $30,000 per year, so why is the lower threshold set at $50,000? And if loans are primarily used by lower income groups, how do people earning $50,000, $80,000, and $110,000 or more fit that profile? 

Surprised by how small the differences are across income groups? 

Me, too. I would have expected to see big differences, especially in loans for discretionary spending on such things as entertainment and travel. But there is virtually no difference between income groups with a range of only 2 percentage points for travel and zero for entertainment. The variance is similarly negligible for car expenses (1), family and children (1), healthcare (3), and loan repayments (1). I would want to know why the differences are so slight.

What do you see when you look at this next table?

The cross-tabs for age show variances of only a few percent for cars, healthcare, loan repayments, and travel. I see two interesting things, though, both involving borrowers under 25. They are roughly twice as likely take out payday loans for entertainment purposes and a great deal less likely to get loans for family and children, the category with the greatest variance (4% to 22%). It makes sense that not many borrowers under 25 are married with kids.

Overall, we are left to wonder how different the numbers would be if someone hadn’t stuffed nearly half of the survey responses into “Other.” The best analysts insist upon having detailed data to look at and design surveys accordingly.

Analytically, I would want to look at some categories of my own

For example, I would want to know more about discretionary spending vs spending on essentials – the difference between things we want and things we need. Unfortunately, we have no good way of determining which is which from the data we are given. It is easy to assume most entertainment is discretionary, but we don’t know if travel means vacations or relocations or commuting, do we? Because we can’t tell what expenses related to families and children means, we can only conclude that some are for essentials and some are for non-essentials – but again we don’t know which. We have no way of knowing which car expenses are essential (the transmission quit working and needs to be replaced) and which are discretionary (sparkly new rims).

About now, inquisitive readers are wondering how much of the “Other” category is discretionary and how much is essential

Oh, by the way, did you notice that there were no categories for food, rent, utilities, cellphones, or public transportation? Surely some of these things are lumped together in “Other,” but which ones and how many?

Every time you see data in infographics, tables, and charts, it is useful to remind yourself you are seeing only the part of the story someone wants to tell you

Why many people really take out payday loans (poor financial skills, short-term priorities) is not the same as the reasons they give for getting them. What people say they do is not the same as what they actually do, is it?

A published analysis of 1,900 payday loan applicants by loans.org paints a different picture

The biggest eyebrow raiser? Nearly one-fourth of applicants borrowed to pay debts. This is ten times the proportion reported in the study we just looked at. If I you paid for the study we’ve been looking at, you would begin to wonder what the reasons might be for such a high level of disagreement between their findings and ours.

• 27% Other
• 24% to pay off debt consolidators.
• 17% household expenses.
• 8% auto.
• 5% home improvement.
• 5% medical/dental.
• 4% vacation.
• 3% wedding loans.
• 2% or less each for large purchases, business, baby and adoption, taxes, motorcycles, and RVs.

I long ago learned that if I can’t find a methods section that shows the sample, the questionnaire, and other important study details, I don’t bother even reading the rest of the report because it will have such little value, for all the reasons we just explored and more. 

Bonus

Whenever I read about payday loans, I think about Wimpy. Popeye’s hamburger-loving friend, J. Wellington Wimpy is always broke and his catchphrase is

The problem is that he never does repay the loan.

Want to read more articles like this? click here