Crunch Time: Fiscal Crises and
the Role of Monetary Policy
David Greenlaw (Morgan Stanley)
James D. Hamilton (UCSD)
Peter Hooper (Deutsche Bank
Securities)
Frederic S. Mishkin (Columbia)
Mechanics of debt accumulation
bt = debt/GDP
st = (primary surplus)/GDP
Rt = nominal interest rate
gt = nominal GDP growth rate
1 + rt = (1 + Rt ) / (1 + gt )  1 + Rt − gt
bt +1 = (1 + rt )(bt − st ).
Fiscal sustainability
r * is constant
Suppose Rt − gt =
*
and current debt/GDP is b . Then holding
debt/GDP constant requires a primary surplus of
∗ ∗
r
b
∗
s =
∗
1+ r
s° = primary surplus that country could plausibly
maintain given its politics and history
What happens if s* > s°?
Tipping points
rt = risk - free net interest rate
e
rt = promised net interest rate
b = sustainable debt level
*
π = Q - measure probability of reform
Q
t
1 + rt = π
e
Q
t (1 + rt ) + (1 −
π
∗
Q
t )
b
.
bt − st
Data used in statistical analysis
• Panel of 20 different advanced countries
• Annual data, 2000-2011
• Question: What factors in year t - 1 help
predict the average yield on 10-year debt
in year t?
Linear regression results
Rit = αˆi + γˆt + 0.0313 bi ,t −1 + 0.0142 bin,t −1 − 0.184 ci ,t −1 + eit
(3.95)
(2.30)
(5.16)
R 2 = 0.69
Rit = 10-year yield for country i year t
bi ,t −1 = previous year's gross debt/GDP
bin,t −1 = previous year's net debt/GDP
ci ,t −1 = previous 5 years' average current account surplus
t statistics in parentheses
Example: if b and bn both increase by one percentage-point
then 10-year yield increases by 4.5 bp
Study
Data
Finding
31 advanced &
emerging economies,
1980-2007
10 advanced
countries, 1990-2010
(Germany only
eurozone)
1% ↑ in debt/GDP → 4
bp ↑ in 5y-10y forward
rate
1% ↑ in govt debt/GDP
plus 1% ↑ in external
debt/GDP → 3 bp ↑ in
5y-10y forward rate
Laubach (2009)
U.S. CBO
projections, 19762006
Reinhart and
Sack (2000)
G7, 1981-2000 (preeuro)
1% ↑ in debt/GDP →
3-4 bp ↑ in 5y-10y
forward rate
1% ↑ in projected
surplus/GDP → 12 bp
↓ in 10 yr - 3 m spread
Baldacci and
Kumar (2010)
Ichiue and
Shimizu (2013)
Nonlinear regression results
Rit = αˆi + γˆt + 0.0029 bi ,t −1 + 0.245 ci ,t −1 + 0.000203 bi2,t −1
(0.30)
(4.29)
(4.81)
+ 0.00793 ci2,t −1 − 0.00636 ci ,t −1bi ,t −1 + eit R 2 = 0.82
(2.98)
(10.18)
Rit = 10-year yield for country i year t
bi ,t −1 = previous year's debt/GDP
ci ,t −1 = previous 5 years' average current account surplus
t statistics in parentheses
Example: if c = 0 and b = 60%, a one percentage-point
increase in b leads to 3 bp increase in 10-year yield
Change in interest rate associated
with higher debt and CA balances
Sovereign yield (%)
16
14
12
CA = 0
10
8
CA = -2.5%
6
CA = -5%
4
2
0
0
20
40
60
80 100 120 140 160 180 200
Debt/GDP (%)
Greece in 2008: debt/GDP = 100%,
R = 5%, g = 6.6%
One measure that actually helped
Greece was PSI default
Case study: Ireland
Index
Jan 2000=100
EUR
200
400000
CoreLogic national house price index(ls)
Ireland: average price of existing homes(rs)
160
300000
120
200000
80
100000
40
0
80
85
90
95
00
05
10
Nationalizing banking losses
pushed Ireland into tipping point
Italy’s growth made 120% debt
sustainable in 1995 but not today
Why is Japanese yield still so low?
• Our regression: favorable country fixed
effect and large current-account surplus
• Why is Japan special? High domestic
saving rate and extreme home bias
• Hoshi and Ito (2013): Japan’s declining
saving rate could force it to turn to
international lenders
• Conclusion: Japan may face big
challenges in future
United States federal government net
interest expense is currently 1.4% of GDP
But interest rates expected to rise
CBO projects interest expense will
exceed defense budget by 2020
Methodology similar to CBO’s previous long-term
projections suggests problem just gets worse
33%
31%
CBO's US Budget Projections (% of GDP)
29%
27%
25%
23%
Expenditures
Net Interest
21%
19%
17%
15%
Revenues
19
90
19
92
19
94
19
96
19
98
20
00
20
02
20
04
20
06
20
08
20
10
20
12
20
14
20
16
20
18
20
20
20
22
20
24
20
26
20
28
20
30
20
32
20
34
20
36
10-year Treasury rate assumed in
previous simulation
15
Projections
12
Actual
9
6
3
0
Gross debt/GDP assumed in
previous simulation
Gross Debt to GDP, Percent
250%
225%
200%
175%
150%
125%
100%
75%
2012 2014 2016 2018 2020 2022 2024 2026 2028 2030 2032 2034 2036
Gross debt with feedback to
interest rate
250%
rising interest rate
225%
constant interest
rate
200%
175%
150%
125%
100%
20
36
20
34
20
32
20
30
20
28
20
26
20
24
20
22
20
20
20
18
20
16
20
14
20
12
75%
Gross debt with sequester cancelled
and 6% steady-state unemployment
350%
300%
6% unemployment
and no sequester
constant interest
rate
250%
200%
150%
100%
20
36
20
34
20
32
20
30
20
28
20
26
20
24
20
22
20
20
20
18
20
16
20
14
20
12
50%
What can central banks do about
this?
• Central bank can use monetary ease to offset
contractionary fiscal policy to make successful
reform more likely
• Central bank cannot bring sustainability to an
otherwise unsustainable fiscal policy
• Replacing long-term Treasury debt with interestpaying reserves shortens maturity structure of
combined Treasury-Fed balance sheet
• Shorter maturity structure makes countries more
vulnerable to tipping-point dynamics
• Not a matter of central bank credibility-Fed has no good options if Treasury
auction undersubscribed
• Mechanical question: what would
consequences of higher interest rates be
for Fed’s balance sheet?
• Analyzed independently by Carpenter,
Ihrig, Klee, Boote, and Quinn (2013)
Variable
Asset
purchases
Asset sales
Assumed growth path
Continue at current pace through December
2013, slow to maintenance levels (stock
stable) through 2014, stop (stock declines)
in 2015.
MBS sales start late 2015, completed in 2019
MBS
Follows market models
prepayment
Liabilities
Interest rates
Currency grows at 7% AR (2pp above Blue
Chip forecast for nominal GDP growth
per historical experience); required
reserves grow at 4% AR
Driven by Blue Chip consensus forecast
Fed capital
Grows at 10% AR per historical average
Operating
Grow on historical trend
Fed Balance Sheet
Assets
$ Bln
4,000
Baseline Projections
3,500
3,000
Total assets
w/ no asset
sales
MBS
2,500
2,000
10+ yr
1,500
5-10 yr
Treasuries
1,000
1-5 yr
500
< 1 yr
0
2004
2006
2008
2010
2012
2014
2016
Sources: FRB, Haver Analytics and authors’ calculations
2018
2020
Fed Balance Sheet
$ Bln
Liabilities
4,000
Baseline Projections
3,500
Excess
Reserves
with no
asset sales
3,000
2,500
2,000
Required
Reserves
Excess
Reserves
1,500
1,000
Currency
500
0
2004
2006
2008
2010
2012
2014
2016
Sources: FRB, Haver Analytics and authors’ calculations
2018
2020
%
Interest rate assumptions: baseline
%
6.0%
6.0%
2017-2020
5.0%
5.0%
2016
4.0%
2015
4.0%
3.0%
2014
3.0%
2013
2.0%
2.0%
1.0%
1.0%
0.0%
0.0%
0
3
6
9
12
15
18
Maturity in Years
21
24
27
30
Fed net income account
$ Bln
$ Bln
Operating expense
Capital surplus plus dividends
120
120
Baseline Projections
100
100
80
80
60
60
Interest income
40
40
20
20
0
0
Interest payments
-20
-40
-60
2005
-20
-40
Realized losses (or gains)
-60
2007
2009
2011
2013
2015
Sources: FRB and authors’ calculations
2017
2019
Net Income Available for
Remittance to Treasury
$ Bln
$ Bln
Projections
120
120
100
100
Baseline
80
60
80
40
No asset 60
sales
40
20
20
0
0
-20
-20
+100 bp risk
premium
-40
-60
2005
2007
2009
2011
2013
2015
Sources: FRB and authors’ calculations
-40
-60
2017
2019
Cumulative net income relative to
pre-crisis trend
$ Bln
$ Bln
Projections
400
400
Baseline
300
300
200
No asset
200
sales
100
100
+100 bp risk
premium
0
-100
2005
0
-100
2007
2009
2011
2013
Sources: Authors’ calculations
2015
2017
2019
Alternative exit and fiscal shock
scenarios
Peak
deferred
asset ($ Bn)
Cumulative
excess
gain/loss ($ Bn)
2007-2020
Baseline
22
43
100 bp risk premium begin 2016
105
-43
200 bp risk premium begin 2016
194
-126
200 bp risk premium, no asset sales
67
-18
QE through 2014 at current pace
58
45
QE through 2014,
200 bp risk premium begin 2016
372
-282
Sources: Authors’ calculations
Conclusions
• Odds of tipping point rise sharply with debt above 100%
and current-account deficit above 2.5%; at these levels,
1% increase in b increases R by 6 basis points
• US debt stabilizing near term, but at levels that put it at
risk; longer-term US debt still on unsustainable path
• A jump in Treasury risk premium in next five years would
complicate Fed exit, resulting in negative income
• Fed credibility undermined and inflation expectations
boosted as pressure builds to shift in an easing direction