For those who keep somewhat up to date with me as a person rather than as a writer, you may have noticed I’ve moved to Australia. I also got a job there that now pays me fortnightly. And that’s exactly what today’s post is about. In the Netherlands and the UK where I’ve lived, studied and worked before I’ve always been on a monthly pay-out cycle. Wage, subsidies, benefits etc. it’s all done on essentially the same date, granted that that date is not a weekend day (banks, am I right?!). If you don’t know any better, a monthly payment cycle is not something you really question. Until you do. I’m now being paid fortnightly, as I’ve mentioned before, and it does seem to have an impact. Not only do you get money twice as often, you also only get half of it per go. Quite the adjustment. I can see my own mental accounting change as a result of it. I’m postponing purchases and working with smaller budgets. Now that’s all well and good for me, but extrapolate this small personal change into the hundreds of thousands or even millions working in the gig economy who don’t have stable streams of incomes and are on “God only knows” types of payment cycles, and you can see why I’m interested in this phenomenon.
There is a paper out there, somewhere, which describes the effect of doing tax rebates in smaller chunks rather than one big rebate. This paper is the type of famous where every one knows it, but no one knows where to find it (sketchy, I know). The gist of this unicorn paper: if you give people tax rebates in smaller chunks (monthly, per quarter etc.) they are more likely to spend this money completely and not save any of it, as compared to receiving a larger chunk of rebate once annually. The idea here is simple: if I give you back $1000 in tax rebates you see it as a large sum of money which can make a difference: save some of it, and spend the rest as a bit of a windfall. This rebate is nice. If I give you back ~$85 ($1000/12) every month, that’s not “enough” money to make a dent, whether that dent is positive (debt repayment, savings) or negative (increased spending at the expense of the former categories). Knowing this, would a different wage payment cycle impact things in the same way? As I mentioned, I’m starting to reframe my approach to budgeting and mental accounting to work with smaller, yet more frequent pay outs. My reasoning, however, is still on a monthly basis: “I make X per month, which is only a payout of .5X every 2 weeks, and I want to save .2X of my wage, which means setting aside .1X every time I get paid.” Work with fractions or work with absolutes, it matters relatively little. Point is, I know what I’m working with, both monthly and fortnightly. Now the main thing I’ve noticed is that I’m postponing larger purchases AND I find myself much more likely to opt-in (or at least consider) using BNPL (buy now pay later) services such as Klarna, PayPal’s in 4 or Afterpay (dangerous, I know). And this is odd, because that’s exactly the opposite of what this amazing paper by Leary & Wang (2016), who find that 1) borrowing is procyclical with liquidity over the pay period, with loan volume declining by 47% over the course of a pay period, and increasing discontinuously on pay days; 2) borrowing per day is 38% higher during 35-day compared with 28-day pay periods (the difference week can make!); and 3) consumers borrow 3% less if they are assigned to receive income on the fourth Wednesday compared to the second Wednesday of the month, consistent with imperfect planning for recurring expenditure commitments at the beginning of the month. But I think, most shockingly and most worryingly, their results suggest that failures to adjust to predictable variation in income timing account for at least 18% of payday loan volume. That’s a huge number. To put that in market cap., in the 2016 market 18% is about $29-41 million in excess costs per year among benefits recipients. So yes, timing matters. So what you’d think would happen is that smaller more frequent, but also predictable payments should help with consumption in a way that it makes budgeting easier, and less necessary to go into suboptimal forms of debt. However, if we extend this back into the rebate literature, that doesn’t seem to hold up: A much older paper by Shapiro & Slemrod (1993) looks at a similar question, once again looking at tax refunds under the Bush policy. The tax refund was a result of a policy change and amounted to approximately $25 per month, and was in fact paid out monthly. They find that when surveying households what they are planning to do with this refund just over half is planning to save it (save + repay debt), which a large minority (48%) planning to spend it fully. It has to be mentioned however, that the idea that $25 is a rather small amount of money, and about one third of those surveyed didn’t actually realise the money had come in when it did as a result of the policy. So my idea of pay-outs under $100 not really making a dent stands. Moving from 1993 to almost 1 decade later, we are still looking predominantly at rebate research. Continued work by Shapiro & Slemrod (2003) finds that in most cases, tax rebates do not really stimulate spending that much, “only 21.8 percent of households report that the income tax rebate of 2001 led them mostly to increase spending,” which means that the remainder is still going into savings/debt repayment, if it is repaid as a lumpsum. Jumping up another few years to 2009, Shapiro & Slemrod find essentially the same pattern all over again, with “only one-fifth of respondents said that the 2008 tax rebates would lead them to mostly increase spending. Almost half said the rebate would mostly lead them to pay off debt, while about a third saying it would lead them mostly to save more.”
Now the study of rebates is not the study of payment cycles and their effect on spending, savings and debt repayment/uptake. So we have to focus mostly on the unicorn paper mentioned in the introduction and the work by Leary & Wang (2016). And what do we have there? Well, an interesting balancing act of more frequent wage pay-outs helping with consumption smoothing, but only to the extent that these payments actually seem to make a dent. And there’s the crux of the matter. Because can you guarantee that gig-workers experience there highly frequent pay-outs as big enough to make a dent that they would consider saving/paying off debt, rather than spending it? People have a tendency to underestimate small amounts, this has been incredibly well-documented in behavioural science, so where does that leave us here? It seems that a monthly payment cycle is not optimal (Leary & Wang, 2016), but gig-work is too little too often, and often too unpredictable to actually aid consumption smoothing. We are finding ourselves at a tipping point in a careful balancing act, with no real conclusions to draw except for this one: we need more research on payment cycles and their effects on personal finance management.
Oh and btw, the unicorn paper I referred to earlier? It’s by Chambers and Spencer (2008) and listed down below. So off reading you go!
Chambers, V., & Spencer, M. (2008). Does changing the timing of a yearly individual tax refund change the amount spent vs. saved?. Journal of Economic Psychology, 29(6), 856-862.
Leary, J., & Wang, J. (2016). Liquidity constraints and budgeting mistakes: Evidence from social security recipients. Unpublished.
Shapiro, M. D., & Slemrod, J. (1993). Consumer response to the timing of income: Evidence from a change in tax withholding. Shapiro, M. D., & Slemrod, J. (2003). Consumer response to tax rebates. American Economic Review, 93