I have been getting versions of this question a lot. It is very hard to answer with any precision, so, below are some very imprecise thoughts.

First, the shutdown would have to go on for quite a long-time (say at least a month) to affect the trajectory of aggregate macroeconomic statistics like gross domestic product (GDP) or employment growth. For one, the majority of what the federal government spends money on (including the health insurance coverage expansions contained in the ACA!) will not be affected by the shutdown. Transfers payments like Social Security, Medicare, Medicaid, Food Stamps, etc…will continue to flow, as will essential discretionary spending.

Given the relatively restricted scope of the shutdown in terms of government spending, it stands to reason that it would have to go on for a a month or so before there would be enough of a mechanical fiscal drag to start significantly affecting the path of macroeconomic aggregates. A very, very rough back-of-the-envelope estimate would be that the strictly mechanical impact of a month of the shutdown would subtract 0.1-0.2 percentage points off of GDP growth for (fiscal) 2014.* So, if the government shutdown lasts a month and the economy was set to grow 3 percent in 2014 without the shutdown, the mechanical drag from the shutdown would result in actual growth of 2.8-2.9 percent. Of course, if one focused on the effect of the shutdown only in the fourth quarter of (calendar) 2013, it will matter quite a bit more (multiply that 0.1-0.2 by 4, so, a one-month shutdown would reduce fourth quarter GDP growth by about 0.4-0.8 percent, which is not peanuts for a quarterly growth number).

But, this talk about measuring the effect on growth quarter-by-quarter leads to a key unknown in just how damaging it will eventually be to the overall trajectory of the recovery: whether or not federal employees receive back-pay and what the funding levels are going forward from the resolution. If employees receive back-pay (which seems like the only morally acceptable thing to do—for whatever that is worth in this debate) and if funding levels for the year are unaffected by the shutdown, the macroeconomic effects will be much smaller than if employees do not receive pay and/or annual funding levels are reduced.

Before I move on, let me make one thing clear—this assessment does absolutely not mean that the shutdown is no big deal. It’s a very big deal—it will harm people who need help from public employees and won’t get it. It will make government less efficient even before and after the shutdown, as now resources will flow to managing this and future potential shutdowns rather than focusing on the core business of government. Further, in the D.C. metropolitan economy, it needn’t be that long a shutdown at all to make a real difference and cause suffering—not just among federal employees but among the shopkeepers and restaurant owners that cater to them. In short, not everything in the world needs to show up clearly in GDP statistics to matter a lot and cause real distress.

The larger macroeconomic danger of the government shutdown for overall recovery from the Great Recession, however, is that its resolution leads to a substantially reduced level of public spending in coming years. As I noted before, the real damage done by past fiscal showdowns has not been the immediate damage caused while they persisted, but was instead the future austerity they locked in. As I’ve been harping on for a while, the current recovery has seen historically low public spending, and this low public spending is entirely why we remain deeply depressed from the Great Recession. So, in this sense, the shutdown is indeed extraordinarily damaging even in very broad macroeconomic terms.

To make this concrete—the silliness over “defunding Obamacare” obscures the fact that even the continuing resolution that would almost surely get Democratic support bakes in the full effect of the sequester for 2014. Repealing the sequester would boost job-growth by nearly a million jobs in 2014, yet we’re not going to do it largely because Democrats will just be eager to get past the immediate threat to the Affordable Care Act (ACA)—and to be clear that’s not necessarily a terrible decision on their part.

I should note that there are smart economists who weigh the danger posed by the government shutdown to consumer confidence and generalized uncertainty much higher than I do, which could amplify the mechanical effects I noted above. Early signs are that consumer confidence may indeed not much be helped by talk of the shutdown. One doesn’t have to be an alarmist to realize that we have a weak, fragile recovery already, so anything that drags on it is awfully dangerous, even if it doesn’t radically change the overall path of recovery.

Finally, it’s worth noting that a prolonged government shutdown would run this particular fiscal showdown right into the next one upcoming on the schedule—the breaching of the statutory debt ceiling in mid-October (or later, given the significantly reduced path of spending implied by a prolonged showdown). It’s granted by one and all that defaulting on outstanding U.S. debt would be extraordinarily damaging from a macroeconomic perspective; it would likely make the financial crisis of late 2008 look tame. But because it would be so damaging, and because there would appear to be ample legal workarounds available to the Obama administration to keep it from coming to pass (yes, the trillion dollar coin is one of them, and simply saying that it sounds silly is not a real argument against it), it would be an incredible failure for this to actually happen.

*Very, very rough numbers are that discretionary spending in total was likely to end up around $1 trillion in 2014. Say that the shutdown reduces this by a third (one-third is roughly the share of non-exempt employees in the federal workforce, so, say that spending is affected at the same rate as government employment). This is $330 billion—or about 2 percent of GDP that would mechanically be sucked out of the economy if the shutdown lasted a full-year. Divide this by 1/12th to get a monthly drag and you get right around 0.1 percent of GDP, and with a generic multiplier attached to it could conceivably push this number up to close to 0.2 percent of GDP.