Tax refunds, which can be an important source of cash flow for consumers early in the year, have totaled $20bn less year-to-date than refunds in 2012. Goldman Sachs notes that this is the equivalent of nearly 1% of disposable income over that period, and some consumer-oriented businesses have attributed lackluster sales in late January and early February to lower refund payments. Balancing the possibility of a small amount of additional catch-up with the possibility that some of the decline versus 2012 is fundamentally driven by the effects of tax law changes or other factors, the upshot is that Goldman believes the cumulative gap of around $20bn looks likely to persist. Since the current rate of change in tax refunds looks similar to last year's, this should not weigh further on consumer cash flow. However, it also implies that we should not expect the consumer to receive much of a tailwind from delayed tax refunds in March or April. It does make one wonder a little if this marginal cash-flow is the reason for the extremely unusual cyclical strength and weakness we have seen in macro data for the last few years.
Via Goldman Sachs,
- Over the first nine weeks of the year, tax refunds have totaled $20bn less than in the same period in 2012. Some of this was clearly due to delays in the tax filing season caused by enactment of tax legislation at the start of the year.
- However, while the decline in tax refunds versus 2012 has partly reversed, most of the catch-up due to the late start to the tax filing season appears to have taken place, and refunds have begun to taper off as they typically do in early March. Balancing the possibility of some additional catch-up with the likelihood that some of the gap is fundamental in nature and could thus widen as the remaining refunds are paid, the gap looks likely to persist at around the current level.
- The upshot is that while weak tax refunds have stabilized and therefore should not weigh further on consumer cashflow, now that refunds have caught up to prior year patterns we would not expect them to provide a tailwind to consumers in March or April, either.
Part of this weakness clearly relates to the delayed start to the tax filing season, which was caused by the enactment of legislation to avert most of the "fiscal cliff" at the start of the year. The result was that taxpayers were unable to file tax returns until the last days of January, which in turn delayed the flow of tax refunds. By the time the Internal Revenue Service (IRS) began to pay refunds at the start of February, cumulative payments were about $12bn behind where they were in 2010, 2011, or 2012. The gap widened a bit further in early February versus the prior year, but has since narrowed and at this point stands at around $20bn (Exhibit 1).
There may be a small amount of additional catch-up left to occur. Some taxpayers with certain types of credits could not file until mid-February and an even smaller group could not file until this week. However, the number of returns that were delayed was probably fairly small. For example, one of the more common items that caused delays was an education credit the IRS says is typically claimed on around 5% of returns received by mid-February, when the IRS started accepting it. So while there might be a small amount of catching up left to do, resolving these additional delays will probably only narrow the gap slightly.
There are other reasons to believe total refunds will remain lower than in 2012. The first is that refunds actually fell versus the prior year in 2011 and 2012 as well, by $7bn and $9bn, respectively. It isn't clear why refunds declined in those years, but a factor affecting the tax filing season now underway may be the realization of capital gains ahead of the year-end increase in tax rates. The Joint Tax Committee estimated in 2012 that activity ahead of the year-end increase in capital gains and dividend tax rates would increase tax revenues by $21bn in 2013, and it is likely that some of this is showing up as a reduction in tax refunds. (Non-withheld tax payments were also up sharply in January, when estimated tax payments were due for Q4 2012, when it is likely that some of this tax-driven selling may have taken place.)
Refund data over the last few days also indicate that payments are beginning to decline. As shown in Exhibit 2, refunds begin to taper off around this time each year, and over the last few days refund payments have followed a very similar path to last year's. This could obviously still change, and a small change in the path of refunds can have a large cumulative effect. For example, if refunds over the next several weeks followed the path in 2011 (a slightly less front-loaded year than 2012), the cumulative gap with 2012 would be reduced to $10bn by the end of May. That said, for now there are few indications that refunds will stay higher for longer than they did last year.
Finally, it is worth noting that at this point in the tax filing season, only about half of tax refunds have typically been paid. In normal years when refunds are growing over the prior year, the gain in dollar terms over the prior year by the end of the tax filing season tends to be around twice as large as the gain half-way through. The relationship is inconsistent when there is a shortfall in refunds versus the prior year (there are only a few examples of this and no particular pattern) but it does raise the risk that with around half of refunds yet to be paid, the gap could widen a bit further.
Balancing the possibility of a small amount of additional catch-up with the possibility that some of the decline versus 2012 is fundamentally driven by the effects of tax law changes or other factors, the upshot is that the cumulative gap of around $20bn looks likely to persist. Since the current rate of change in tax refunds looks similar to last year's, this should not weigh further on consumer cash flow. However, it also implies that we should not expect the consumer to receive much of a tailwind from delayed tax refunds in March or April.


