Enactment of the American Tax Relief Act of 2012 (ATRA) eliminated
the near-term uncertainty surrounding tax policy, shifting attention to three
other downside fiscal risks to our forecast.
- A brief encounter with the debt ceiling this winter might not be calamitous, but a prolonged one would cause the deepest recession since the Great Depression.
- Just the possibility of hitting the debt ceiling could create enough uncertainty to reduce GDP by roughly ½ percentage point this year, even if the ceiling is raised before it is reached.
- Imposition of the full sequester on March 1 would reduce GDP growth by roughly ¾ percentage point in 2013 relative to our forecast.
- A brief shutdown in April of the federal government would cut GDP growth in the second quarter by about ¼ percentage point for each week it lasted. However, the longer the shutdown, the worse and more chaotic the outcome.
Three Looming
Deadlines
ATRA may have clarified near-term tax policy but three
looming fiscal deadlines pose downside risks to our forecast. First, unless the debt ceiling is raised, the
Treasury will be unable to pay all its bills by late February or early
March. Second, ATRA delayed until March
1 the sequester originally enacted under the Budget Control Act of 2011 (BCA). Our forecast presumes the full sequester is
avoided, replaced instead with smaller spending cuts, but we could be
wrong. Finally, the federal government
is operating under a continuing resolution that expires March 27. Unless new spending authority is legislated
by then, non-essential government functions could be shut down at the end of
the first quarter.
Debt
Ceiling
The United States hit its debt ceiling on December 31,
2012. The Treasury will use temporary
financial maneuvers to continue paying its bills through late February or early
March.[1] Thereafter, unless the debt ceiling is raised
— or some other strategy is adopted to circumvent it[2]
— the federal government will be unable to meet all of its current financial
commitments in a full or timely manner.
In that event, the Treasury likely would manage its cash flow to meet
interest payments on the national debt, thereby avoiding a sovereign default.[3] However, the government could fall into
default on contracts with private vendors, and be perceived as having broken
long-standing social compacts with the beneficiaries of entitlement programs.
A De Facto Balanced
Budget Amendment
In terms of expenditure flows, hitting the debt ceiling
is like enforcing a Balanced Budget Amendment (BBA) daily and from the spending
side of the ledger only. While the
Treasury could meet all its interest obligations, the budget would have to be
brought into balance immediately and kept there by varying other outlays. The result would be a reduction in the level
of spending that, on average, would be not only very large, but also, given the
seasonal “lumpiness” of tax receipts, very volatile.
Some have suggested that if the Treasury avoids sovereign
default, damage from enforcing the debt ceiling for even a lengthy period could
be limited by prioritizing and juggling other payments. To appreciate the folly of this view,
suppose the debt ceiling became binding in March and remained so through 2014. For the rest of 2013, the “static” reduction
in spending would average roughly $825 billion at an annual rate, or roughly 5%
of GDP.[4] This is a huge fiscal contraction to occur
abruptly. For example, we recently
estimated that the contraction associated with the “fiscal cliff,” which was a
smaller 4¼% of GDP, would have thrown the economy into recession. Furthermore, most of the contraction
associated with the fiscal cliff was from tax increases, which have smaller
multipliers than the spending cuts necessary to enforce the debt limit. And, were such cuts implemented for a
sustained period, as the economy fell into recession and tax revenues declined,
additional cuts would be required to keep the budget balanced and avoid
additional debt issuance. This, in turn,
would further steepen the economic downturn.[5]
Just how dire might be the outcome? To see, we ran a simulation in which the debt
ceiling becomes binding in March and remains in force through the end of
2014. The Treasury is assumed to meet
all interest payments on the national debt but cut other spending
proportionately to bring the budget into continuous balance through the fourth
quarter.
For example, our baseline forecast shows a unified budget
deficit in the first quarter of $1.095 trillion at an annual rate (not
seasonally adjusted). However, in three
of the last four years, roughly 40% of first-quarter deficits have accumulated
during March. If that same pattern were
to repeat this year, it would take a static spending cut of $1.314 trillion
against our baseline to balance the budget in March, thereby reducing first-quarter
spending by $438 billion (= 1/3 of $1.314 trillion), all at annual rates. For the second quarter, our forecast shows a relatively
small deficit of $533 billion because April sees strong revenues. This requires a sequential cut in spending
from the first to the second quarter of $95 billion (= $533 billion less $438
billion, but implying a huge $800 billion rebound in spending from the March
level). In the third quarter, our
baseline deficit rises to $745 billion, requiring a sequential cut in spending
of $213 billion (= $745 billion less $532 billion), and the respective numbers
for the fourth quarter are $809 billion and $64 billion (= $809 billion less
$745 billion). In the simulation, even
larger “dynamic” cuts in spending are required to offset the loss in revenue
associated with the weakening economy.
This pattern of expenditure cuts is repeated in 2014, but
that can’t be the whole story. Appetite
for risk would also decline in this scenario, although it is hard to know by how
much, and for how long. However, some allowance must be made for the
uncertainty that would accompany 30% absolute
(that is, not relative to a rising baseline) across-the-board cuts in
non-interest outlays against the backdrop of abrogated legal contracts and
social compacts. In our model, the VIX
is used as a general marker for risk that influences the prices of risky assets
which, in turn, affect aggregate demand.
For this experiment, we introduced a spike in the VIX two-thirds the
size, but of the same duration, as the spike that occurred during the financial
crisis in 2008 and 2009.
The results of a model simulation involving both
the required spending cuts and the spike in the VIX are shown in Charts 1 and
2.
The unified budget deficit falls into
balance after the first quarter of this year.
GDP growth, shown here as the percent change from four quarters earlier,
turns sharply negative, hitting -8% in the fourth quarter of this year before
then rising to 0 by the end of 2014. This would be the worst recession since
World War II by a wide margin, sending the unemployment rate soaring above 14%
by the middle of next year. We don’t
want to make too much of this simulation, since even in today’s rancorous
political climate such a disastrous scenario seems improbable. The results are more relevant for a
discussion about the impact of a strict balanced budget amendment. Nevertheless, they do emphasize that if the
Treasury is constrained by the debt ceiling for any meaningful period of time,
economic performance could worsen quickly and quite dramatically.
Uncertainty is Bad
Enough
Perhaps the most likely scenario is that during the
exercise of partisan brinkmanship, the US approaches or even briefly hits the
debt ceiling and that, given its daily cash flows, the Treasury has
insufficient balances to meet a contractual obligation or pay promised benefits
on time, thereby “defaulting” on those commitments before a resolution to the
political impasse is reached. This
scenario is not as dire as a sovereign default, but it is bad enough since it
would heighten uncertainty about federal financial flows, call into question
the meaning of “full faith and credit” of the U.S. government, undermine the
credibility of our political institutions, reduce our standing with the rest of
the world, and encourage rating agencies to downgrade U.S. debt further.
The events of the summer of 2011 suggest that the
political brinkmanship preceding the eventual resolution of the debt-ceiling
crisis in August created uncertainty that slowed GDP growth in the third
quarter of that year. We could be in for
a repeat of those events during the first quarter of 2013. How much might such uncertainty slow growth
in 2013?
To answer this question we began by examining the
Index of Policy Uncertainty (IPU) developed by Baker, Bloom, and Davis.[6] The IPU reflects mentions by the media of
policy uncertainty, forecasters’ disagreement about federal purchases of goods
and services, forecasters’ disagreement about inflation, and scheduled
expirations of taxes. The index spiked
in the summer of 2011 during the dispute over the debt ceiling. (See Chart
3). Intuitively, such uncertainty could
slow economic growth by encouraging households and businesses to delay
decisions and by reducing the price of risk assets.
The IPU does not appear in our macro model, but Baker et.
al. emphasize that their index is correlated with the VIX, which, as already
noted, functions in our model as a general marker for uncertainty. Hence, we developed a simple model relating
the VIX to the IPU, and, using that equation, we estimated that the spike in
the IPU during the last debt-ceiling crisis raised the VIX by about 4½ points
during the three months centered on August of 2011. [7],[8] Our published forecast does not show a VIX
shock in the first quarter related to the uncertainty over the debt ceiling. However, by introducing an uncertainty shock
similar to the one associated with the debt-ceiling debacle of 2011, we can
simulate the potential impact on GDP growth of a similar stand-off in 2013.
We ran such a simulation assuming that a 3-month shock is
centered on February, when the debt ceiling could first become binding. The results of this experiment are shown in
Chart 4. They suggest that a temporary
rise in uncertainty surrounding the approach to the debt ceiling could trim 0.4
percentage point from GDP growth over the course of 2013.
Sequester
ATRA delayed the BCA sequester by two months (to March 1)
and paid for the delay by reducing in 2013 and 2014 the caps on discretionary
spending implemented under the Budget Control Act of 2011. While our forecast shows a path for
discretionary spending that is lower than
implied by these modified BCA spending caps, it is above the path
implied by the now-delayed sequester.
(See Chart 5.)
Suppose our
assumption that the full sequester is avoided proves wrong, and on March 1
discretionary spending drops from our baseline to the lower path consistent
with ATRA. How would that affect our
projections for growth this year and next?
We simulated this as a reduction in the level of
federal consumption and gross investment (relative to our forecast) that builds
from $4 billion in the first quarter of this year to $65 billion by the fourth
quarter, and then is maintained through 2014.
In addition, the sequester would reduce the level of spending on
healthcare entitlements by an amount that grows to roughly $10 billion per
year. The results of this simulation for
GDP growth are shown in Chart 6, and the effects are sizable.
GDP growth is reduced by 0.7 percentage point
over the four quarters of 2013 (from 2.6 to 1.9) and by nearly a full
percentage point over the second half of the year as the cuts build
cumulatively. The static drag ends in
2014, but, given lags and multiplier effects, GDP growth in 2014 is only
marginally higher than in our forecast.
Shutdown
The continuing resolution under which the federal
government is now operating expires on March 27. Unless that authority is extended,
non-essential functions of government will shut down, as was threatened in 2011
and as occurred twice during the budget wars between President Clinton and then
Speaker of the House Newt Gingrich. A
federal shutdown seems a real possibility.
High-ranking Republicans are regularly quoted threatening such action.
In early 2011 we published a detailed analysis of how a
brief shutdown of the federal government would be treated in the National
Income and Product Accounts and hence how it would affect GDP growth.[9] It was a quick exercise to update that
analysis.[10] If, starting on April 1, a shutdown began
affecting 36% of federal civilian workers, the static effect would be to reduce GDP growth in the second
quarter by roughly ¼ percentage point for each week the shutdown lasted, and
raise it a similar amount in the third quarter.[11] The effect would arise mostly from the
furloughing of non-essential federal workers whose production is valued at
labor costs in the National Income and Product Accounts. If the shutdown was (expected to be) brief,
the knock-on effects would be small.
However, the longer the shutdown lasts, the broader its impact becomes
and the more one must factor into the analysis multiplier effects as well as
the escalation of uncertainty that would accompany a sustained disruption in
government services and payments to federal contractors.[12] Of course, if the shutdown persisted long
enough it could become quite severe because, unlike the de-facto BBA — which
merely reduced spending to the level of taxes — a prolonged shutdown of the
federal government could eventually zero
out many categories of expenditures.
Summary
Remarks
Fiscal policy is a mess.
Both Republicans and Democrats involved in budget negotiations maintain
positions hardened by ideological differences over the role of government in
society, differing views about the trade-off between the short-run costs of a
fiscal contraction and the long-run benefits, and political positioning.
In our view, a sharp fiscal contraction — or even the
threat of one — imposed when the economy is still struggling to recover from
the great recession and when the Federal Open Market Committee has limited
options to offset any new fiscal drag, is bad policy. It is made worse if implemented outside the
normal budget process in a series of stop-gap actions that, by using calendar
considerations as points of leverage, create attendant uncertainty.
The debt ceiling, the sequester, and the expiring budget
resolution comprise a three-pack of uncertainties that in the near term is bad
for the economy. Because we assume a
benign and sensible resolution to the current fiscal dispute in relatively
short order, at least the fiscal risks around our forecast appear asymmetric to
the downside.
[1]
These maneuvers, which include delaying payments to federal pension funds,
create about $200 billion of headroom for the Treasury to operate early in the
year.
[2]
Three such strategies have been discussed in the media. In the first, the President, citing Section 4
of the 14th Amendment to the Constitution which reads “the validity
of the Public Debt of the United States, authorized by law…shall not be
questioned,” simply ignores the debt limit and instructs the Treasury to
continue issuing debt. In the second,
the Treasury uses an obscure legal authority to mint a platinum coin with a $1
trillion denomination, deposits the coin in the Treasury’s account at the Fed,
and draws on the account to keep paying its bills. In the third, the Treasury issues IOUs —
sometimes referred to as scrips —
that, while technically not debt, serve the same purpose. A secondary market for the scrips could even be encouraged in the
private sector, allowing holders to sell their IOUs at a discount rather than
waiting for the Treasury to honor the pledge later when a deal on the debt
ceiling is finalized. These schemes are certainly inventive, but an attempt to
implement any one of them would almost certainly trigger legal challenges if
not a constitutional crisis. We see all
of them as unlikely and, indeed, the Obama Administration already has ruled out
the first two.
[3]
There has been considerable speculation about how the Treasury might, in the
event of a cash shortfall, prioritize the roughly 80 million payments it makes
every month. Interest payments on the
debt are made over the Fed wire system and could be handled separately with
relative ease. Regarding other payments:
in 2011 the Treasury leaned towards a strategy of simply delaying payments and
then, upon receiving revenues, making what payments possible in the order
due.
[4]
Our current forecast shows NIPA federal net borrowing at $843 billion in
calendar year 2013, relative to nominal GDP of $16,299 billion.
[5]
“Balanced Budget Amendment: A Poor Idea,” Macroeconomic Advisers’ Macro Musing (Volume 4, Number 15; July
29, 2011).
[6]
Scott R. Baker, Nicholas Bloom, and Steven J. Davis, “Measuring Economic Policy
Uncertainty;” www.policyuncertainty.com.
[7]
Not all of the spike in either the IPU or the VIX during the summer of 2011 can
or should be attributed to the policy uncertainty surrounding the dispute over
the debt ceiling. A significant increase
in the perceived risks surrounding the Eurozone crisis also occurred at that
time.
[8]
The equation regressed the change in the monthly average VIX on the current and
lagged change in the Index of Policy Uncertainty (t-statistics in parentheses):
[9]
“From Showdown to Shutdown? The GDP Effects of a (Brief) Federal Government
Shutdown,” Macroeconomic Advisers’ Macro Focus
(Volume 6, Number 2; February 25, 2011)
[10]
Our updated “Shutdown Calculator” is available upon request.
[11]
The first of the two shutdowns during the Clinton-Gingrich standoffs lasted
five days (November 14 to November 19, 1995) and affected roughly 747,000
civilian workers, or one-third the total.
The second shutdown lasted twenty one days (December 16, 1995 to January
6, 1996) but affected only 284,00 civilian workers (or 14% percent of the
total) because by then a defense bill had been enacted covering the civilian
defense employees. The latter
circumstance could be relevant today as both the House and the Senate have
passed defense appropriation bills for 2013.
Hence, it seems likely the Department of Defense will be funded by March
27, limiting the effect of the shutdown to non-defense functions.
[12]
Because they originate from trust funds, Social Security and Medicare benefits
likely would not be threatened.
Contact Macroeconomic Advisers
