Low Equilibrium Real Rates, Financial Crisis, and Secular Stagnation
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
CHAPTER 2
Low Equilibrium Real Rates,
Financial Crisis, and Secular
Stagnation
Lawrence H. Summers
T
he past decade has been a tumultuous one for the US economy,
characterized by the buildup of huge excesses in financial markets
during the 2001–2007 period; the Great Recession and its contain-
ment; and, finally, a recovery that has been very slow by historical stan-
dards and insufficient to bring the economy back even close to the levels
of output that were anticipated before the recession. The containment of
the Recession was no easy feat, since economic conditions initially looked
worse than in the early months of the Great Depression. However, the
economy is still struggling five years later, and the correct diagnosis of its
ailment is requisite for applying the appropriate treatment going forward.
Hence, in this paper I will therefore discuss what I label the new sec-
ular stagnation hypothesis. This hypothesis asserts that the economy as
currently structured is not capable of achieving satisfactory growth and
stable financial conditions simultaneously. The zero lower bound on
base nominal interest rates, in conjunction with low inflation, makes the
achievement of sufficient demand to bring about full employment prob-
lematic. If and when ways can be found to generate sufficient demand,
they will likely be associated with unsustainable financial conditions.
Secular stagnation was first suggested by Alvin Hansen in the late 1930s,1
but did not prove relevant given the rise in demand due to World War II
and the massive pent-up demand for consumer and investment goods
after the war. The difficulty that the US economy has had for many
years in simultaneously achieving full employment, strong growth, and
I am indebted to Simon Hilpert for extensive and excellent assistance in turn-
ing my conference presentation into the current paper.
1. Alvin H. Hansen, “Economic Progress and Declining Population Growth,”
American Economic Review 29, no. 1 (1939): 1–15.
Copyright © 2014 by the Board of Trustees of the Leland Stanford Junior University. All rights reserved.38 Lawrence H. SummerS
financial stability suggests that secular stagnation should be considered
anew. Moreover, the problems of achieving sufficient demand appear to
be even more serious in Europe and Japan than in the United States. I
will argue that secular stagnation is a scenario supported by both theory
and evidence, and therefore is an important contingency to be ensured
against. I will also discuss the policy approaches that could raise demand
and thus help avoid stagnation woes.
Economic facts and a hypothesis
Any explanation of US economic developments in the years leading up to
the Great Recession of 2007–2008, and the five years since, has to grap-
ple with two important facts. First, prior to the crisis, the economy grew
only at a moderate rate and did not overheat. The unemployment rate
stayed above 4 percent and did not plummet to historic depths (figure 2.1).
Similarly, capital utilization did not rise to historically unusual levels
(figure 2.2) and there were no reports of significant shortages in labor
markets. This is remarkable, since multiple factors combined to substan-
tially boost aggregate demand: monetary policy kept interest rates low,
the absence of effective action by financial regulatory authorities and a
breakdown of risk controls brought about excessive leverage in the finan-
cial sector, and the housing markets were characterized by the presence
of large and manifestly unsustainable asset bubbles.
Second, even after the financial system was repaired, the real economy
did not pick up, and growth remained sluggish. The LIBOR OIS spread
(London Interbank Offered Rate, Overnight Indexed Swap), a proxy
for financial distress, was reduced to regular pre-crisis levels by 2009
(figure 2.3); credit default swaps on major financial institutions, which
measure the costs of insuring against a default, quickly normalized (fig-
ure 2.4); and taxpayer funds outstanding to major financial institutions
were largely repaid by the end of 2010 (figure 2.5). All three factors reflect
the swift and successful containment of the crisis and a substantial nor-
malization of conditions in the financial sector. Nonetheless, the broader
economy has not returned to normal—the recovery has only kept up with
population growth and normal productivity growth, but it has not pro-
duced the catch-up growth required to reach the economy’s potential.
Figure 2.6 shows the path of actual GDP (gross domestic product) and
potential GDP as predicted in 2007. With GDP growth limited since the
aftermath of the crisis, there has been almost no gain in output relative
Copyright © 2014 by the Board of Trustees of the Leland Stanford Junior University. All rights reserved.Low Equilibrium Real Rates, Financial Crisis, and Secular Stagnation 39
12.0
10.0
8.0
6.0
4.0
2.0
0.0
ch 8
ay 1
ly 4
pt 7
ov 0
Ja 963
ch 7
ay 0
ly 3
pt 6
ov 9
M n1 2
ch 6
ay 9
Ju 992
pt 5
ov 8
M n2 1
ch 5
11
ay 8
ly 1
ar 94
M 195
Ju 95
Se 195
N 196
ar 96
M 197
Ju 97
Se 197
N 197
Ja 98
ar 98
M 198
Se 199
N 199
Ja 00
ar 00
M 200
Ju 201
20
M n1
1
1
M n1
1
1
1
2
ly
Ja
FIGURE 2.1 Civilian unemployment rate
Source: US Department of Labor, Bureau of Labor Statistics
95.0
90.0
85.0
80.0
75.0
70.0
65.0
ril 67
ly 9
ct 1
Ja 973
ril 6
ly 8
ct 0
Ja 982
ril 85
ly 7
ct 9
Ja 991
ril 94
ly 6
ct 8
Ja 000
ril 03
ly 5
ct 7
Ja 009
ril 12
12
Ju 96
O 197
Ap 197
Ju 197
O 98
Ju 98
O 198
Ju 99
O 199
Ju 00
O 200
Ap 19
Ap 19
Ap 19
Ap 20
Ap 20
20
1
1
1
1
1
1
1
2
2
2
n
n
n
n
n
n
Ja
FIGURE 2.2 Total capital utilization
Source: Board of Governors of the Federal Reserve System
to the previously predicted potential. In the labor market, very limited
progress in restoring the employment ratio (share of the adult popula-
tion that is working) to pre-crisis levels has been made (figure 2.7), even
adjusting for demographic changes (figure 2.8).
The sluggishness of the recovery is counterintuitive on the theory that
the root cause of the output downturn was the financial breakdown in the
fall of 2008. As an analogy, consider episodes characterized by a telephone
Copyright © 2014 by the Board of Trustees of the Leland Stanford Junior University. All rights reserved.40 Lawrence H. SummerS
4
3.5 3m LIBOR-OIS spread
1m LIBOR-OIS spread
3
2.5
2
1.5
1
0.5
0
6
9
0
1
3
2
3
4
5
7
8
2
/0
/0
/1
/1
/1
/0
/0
/0
/0
/0
/0
/1
/4
/4
/4
/4
/4
/4
/4
/4
/4
/4
/4
/4
12
12
12
12
12
12
12
12
12
12
12
12
–0.5
FIGURE 2.3 LIBOR-OIS spread
Source: Bloomberg
connection problem, a power failure, or a breakdown of the transporta-
tion system. While GDP would plummet during such episodes, after tele-
phone connectivity was restored, power turned back on, or transportation
restarted, we would expect the path of GDP to return to normal. For a
time, GDP would be above normal as inventories were replenished and
people caught up with the spending they were unable to do during the
period of failure. However, this has not proved the case for the “financial
power failure” of 2007–2008: now that the central connections have been
repaired, there has been no sign of catch-up, abnormally rapid growth,
or a closing of slack.
The point here may be put starkly. It has been over fifteen years since
the US economy achieved satisfactory and sustainable growth. The Great
Recession of 2007–2008 and subsequent slow recovery followed the
2003–2007 period, which was characterized by bubbles in the financial
markets. Before that was the 2001 recession, which was preceded by the
Internet bubble.
In Japan, it has been a generation since growth approached the 3 per-
cent level that was thought of as a conservative estimate of its potential
Copyright © 2014 by the Board of Trustees of the Leland Stanford Junior University. All rights reserved.4000
3500 JPMorgan Chase
Goldman Sachs
Bank of America
3000
Morgan Stanley
Citigroup
2500 AIG
Wells Fargo
GE Capital
2000
1500
1000
500
0
05
10
11
02
03
06
13
01
04
09
12
07
08
3/
3/
3/
3/
3/
3/
3/
3/
3/
3/
3/
3/
3/
8/
8/
8/
8/
8/
8/
8/
8/
8/
8/
8/
8/
8/
FIGURE 2.4 CDS spreads (bp) on major financial institutions
Source: Bloomberg. Eight largest holding companies as of December 31, 2013
https://www.ffiec.gov/nicpubweb/nicweb/top50form.aspx.
$300
EESA becomes law on Oct. 3, 2008
$250
Dollar Amounts (Billions)
$200
$150
$100
$50
$0
2008 2009 2010 2011 2012 2013 2014
FIGURE 2.5 TARP repayment
Source: US Department of the Treasury, TARP Tracker
Copyright © 2014 by the Board of Trustees of the Leland Stanford Junior University. All rights reserved.22000.0
Actual GDP
Potential GDP, 2007 Estimate
20000.0 Potential GDP, 2013 Estimate
18000.0
16000.0
14000.0
Q1
Q1
20 1
Q1
Q1
Q1
Q1
Q1
Q1
Q1
Q1
Q1
Q
11
16
17
18
07
10
08
13
09
12
14
15
20
20
20
20
20
20
20
20
20
20
20
FIGURE 2.6 Output gap (billions, 2013 USD)
Source: Congressional Budget Office
66.0
64.0
62.0
60.0
58.0
56.0
54.0
ch 7
ov 0
ch 8
ay 1
ly 4
pt 7
M n1 3
ay 0
Ju 973
pt 6
ov 9
M n1 2
ch 6
ay 9
Ju 992
pt 5
ov 8
Ja 001
ch 5
ay 8
ly 1
11
ar 96
N 196
ar 94
M 195
Ju 95
Se 195
Ja 96
M 197
Se 197
N 197
Ja 98
ar 98
M 198
Se 199
N 199
ar 00
M 200
Ju 201
20
M n1
1
1
1
1
1
2
M n2
ly
ly
Ja
FIGURE 2.7 Civilian employment ratio
Source: US Department of Labor, Bureau of Labor Statistics
Copyright © 2014 by the Board of Trustees of the Leland Stanford Junior University. All rights reserved.Low Equilibrium Real Rates, Financial Crisis, and Secular Stagnation 43
96.0
94.0
92.0
90.0
88.0
86.0
84.0
82.0
80.0
55
Ja 58
Ja 6 1
Ja 6 4
Ja 67
Ja 7 0
Ja 973
Ja 7 6
Ja 79
Ja 982
Ja 985
Ja 8 8
Ja 9 1
Ja 994
Ja 9 7
Ja 00
Ja 03
Ja 06
Ja 009
12
19
19
19
19
19
19
19
19
19
19
19
20
20
20
20
1
1
1
1
2
n
n
n
n
n
n
n
n
n
n
n
n
n
n
n
n
n
n
n
n
Ja
Ja
FIGURE 2.8 Employment ratio for men ages 25–54
Source: Organisation for Economic Co-Operation and Development
throughout the 1980s. And the European economy, like the American
economy, had unsustainable finance in the pre-2008 period and has had
manifestly unsatisfactory growth in output and employment since that
time (figure 2.9).
Modern macroeconomics in either its New Keynesian or New Classi-
cal version cannot provide a satisfactory account of this situation. First,
it is a premise of standard formulations of both schools of thought that
fluctuations are cyclical around a path of what is labeled as normal or
trend or potential GDP, so that shortfalls of output in one period are on
average matched by excesses of output in another. In New Classical mod-
els, the fluctuations are frequently seen as optimal responses to changing
economic conditions. In New Keynesian models, fluctuations are treated
as undesirable, but policy can only aspire to reduce the variance of output
over time, not to raise its average level. Clearly, what is required to account
for the experience of recent years in the industrialized world is a theory of
why output is continually depressed relative to potential for a protracted
period. Models that see stabilization policy as an exercise in minimizing
the amplitude of fluctuations around a given mean have little to contribute
to explaining a prolonged period of stagnation a fortiori, as do models
which presume the optimality of outcomes.
Second, models in the dominant macroeconomic traditions attribute
adverse outcomes to some form of wage or price rigidity. In recent years,
the industrial world has been below target inflation despite depressed
Copyright © 2014 by the Board of Trustees of the Leland Stanford Junior University. All rights reserved.44 Lawrence H. SummerS
6 6
4 4
2 2
0 0
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
Average
–2
United States
–4 Euro Area
Japan
–6
FIGURE 2.9 Economic growth in the US, Euro Area, and Japan
Note: The Euro Area contains the fifteen OECD countries that are members of the euro area
(Austria, Belgium, Estonia, Finland, France, Germany, Greece, Ireland, Italy, Luxembourg,
Netherlands, Portugal, Spain, Slovak Republic, Slovenia).
Source: OECD Economic Outlook 2013 (projections for 2013).
levels of output and employment. Greater flexibility of wages and prices
would have exacerbated the situation both by further reducing inflation
and by raising real interest rates—thereby depressing output. So, if any-
thing, in the current context wage and price rigidity are sources of stability
rather than fluctuation.
The secular stagnation hypothesis
The tentative hypothesis of secular stagnation provides an explanation
for the growth patterns just described. Suppose that structural changes in
the US and global economy led to a substantial increase in the propen-
sity to save and a substantial reduction in the propensity to spend and
invest. Then, the real interest rate as the price on saving is supposed to
fall until supply and demand equilibrate. However, short-term safe inter-
est rates cannot fall below zero because people would substitute holding
currency for holding debt instruments that pay a negative yield. Since
instruments that carry risk have a spread beyond safe instruments and
since longer-duration debt has higher yields than shorter-duration debt,
a zero bound on safe government short-term rates implies a lower bound,
albeit above zero, on a broad range of other interest rates that are relevant
for firms’ and households’ decisions.
Copyright © 2014 by the Board of Trustees of the Leland Stanford Junior University. All rights reserved.Low Equilibrium Real Rates, Financial Crisis, and Secular Stagnation 45
With a lower bound on nominal interest rates, savings and investment
cannot be equated by the price channel and hence must be equated by
a reduction in output. This view explains both facts: if equilibrium real
interest rates were low or negative, then the economy would fail to over-
heat or contain significant slack prior to the downturn even with artifi-
cially inflated demand. Also, if equilibrium real interest rates were not
attainable following the crisis, then full employment would not material-
ize even after the financial system was repaired.
Elsewhere I have presented2 a variety of reasons for believing that the
equilibrium real interest rates have fallen. These include (1) decreasing
population growth and possibly declining technological change; (2) an
increase in the tendency to save associated with changes in the income
distribution, with more income being retained by corporations, going to
owners of capital, and going to those with higher incomes and presumably
low propensity to consume; (3) a reduction in the demand for invest-
ment associated with technological changes that reduce the level of capital
investment necessary to carry out a given quantum of economic activity;
and (4) the accumulation of substantial holdings of liquid government
debt by emerging market central banks.
Responding to secular stagnation doubts
Concerns about the possibility of secular stagnation have profound policy
implications. Before analyzing candidate policy responses, two substantial
doubts about the accuracy of the secular stagnation diagnosis need to be
addressed.
First, is growth about to accelerate in the United States and much of
the industrialized world? After all, fears of secular stagnation when raised
in the 1930s were proven wrong. If acceleration is imminent, there is lit-
tle need for great alarm about secular stagnation. Some recent economic
indicators—like the strength of the stock markets and the end of the sharp
fiscal contraction—provide grounds for optimism. However, consensus
forecasts have predicted that escape velocity would be around the cor-
ner for several years, but have been belied by lingering stagnation. Addi-
2. Lawrence H. Summers, “U.S. Economic Prospects: Secular Stagnation, Hys-
teresis, and the Zero Lower Bound,” keynote address at the National Association
for Business Economics conference, February 24, 2014; forthcoming in Business
Economics (2014).
Copyright © 2014 by the Board of Trustees of the Leland Stanford Junior University. All rights reserved.46 Lawrence H. SummerS
tionally, Japan failed to achieve its predicted escape velocity in the 1990s,
despite implementing a zero interest rates following its financial crisis in
the late 1980s; now, Japan has a level of output only a little more than half
of what was forecast twenty years ago. Moreover, even if the economy
accelerates, this provides no assurance that prolonged growth at regu-
lar real interest rates is possible. Across the industrial world, inflation is
below target levels and shows no sign of picking up—a strong suggestion
of a chronic and substantial shortfall in demand.
Second, and related, why should we believe that the economy will not
return to normal levels of output and capacity without additional uncon-
ventional policy? Have real interest rates really declined so substantially
that the zero bound is much more relevant than in the past? As just noted,
there are a range of factors that would suggest substantial declines in equi-
librium real interest rates.
Beyond these theoretical factors, empirical evidence also indicates
a reduction in the real rate of interest. The interest rate on Treasury
Inflation-Protected Securities (TIPS), which is a measure of the real inter-
est rate, has been declining since mid-2007, with the exception of a single
spike in early 2009. Figure 2.10 shows the interest rates on five-, ten-,
twenty-, and thirty-year TIPS. The five- and ten-year real interest rates
have been negative for substantial periods between 2011 and 2013, with the
five-year rate dropping to as low as –1.67 percent in September 2012. Even
the twenty-year TIPS rate dropped into negative territory for a prolonged
period in the latter half of 2012.
A related analysis has been performed by Thomas Laubach and
John C. Williams,3 who seek to determine the equilibrium real interest
rate for the US economy by using sophisticated statistical techniques to
estimate the real interest rate necessary for demand and potential supply
to be equated. Their calculations (depicted in figure 2.11) suggest that
equilibrium real rates are now negative and have been trending downward
for a long time.
Both the abundance of theoretical reasons for a decrease in demand
and the available empirical evidence thus indicate that the US economy is
plausibly incapable of generating demand sufficient to exhaust potential
output, corroborating secular stagnation concerns.
3. Thomas Laubach and John C. Williams, “Measuring the Natural Rate of
Interest,” Review of Economics and Statistics 85, no. 4 (November 2003): 1063–1070.
Copyright © 2014 by the Board of Trustees of the Leland Stanford Junior University. All rights reserved.5
5-Year
4 10-Year
20-Year
30-Year
3
2
1
0
1/2/03 1/2/04 1/2/05 1/2/06 1/2/07 1/2/08 1/2/09 1/2/10 1/2/11 1/2/12 1/2/13 1/2/14
–1
–2
FIGURE 2.10 Real interest rates as measured by returns on
Treasury Inflation-Protected Securities (TIPS).
Source: Board of Governors of the Federal Reserve System
6
4
Percent
2
0
1960q1 1970q1 1980q1 1990q1 2000q1 2010q1
Date
FIGURE 2.11 Estimate for the real interest rate by Laubach and
Williams (2003)
Source: Updated estimates from John Williams’s home page, http://www.frbsf.org
/economic-research/economists/john-williams/.
Copyright © 2014 by the Board of Trustees of the Leland Stanford Junior University. All rights reserved.48 Lawrence H. SummerS
Policy prescriptions
Given the concern that demand is constrained by the lower bound on the
nominal interest rate, there are three potential policy approaches. The first
and least satisfactory is passivity. Perhaps the situation will right itself or
policies that promise an increased emphasis on long-run macroeconomic
rectitude will improve matters. But there is little evidence anywhere in
the industrialized world that such policies in the face of liquidity trap
conditions are availing. Early work suggesting the efficacy of fiscal con-
solidations in stimulating economic activity has been convincingly dis-
credited by research at the International Monetary Fund (IMF) and in
other places.4
The second alternative is to use monetary policy to engineer lower
real interest rates consistent with the zero lower bound on the nominal
interest rate. This involves keeping the federal funds rate near zero and
taking unconventional monetary policy actions that aim aggressively at
reducing risk and term premiums, so that the economically important
risky or longer-term interest rates can be reduced. This strategy is more
attractive than doing nothing, but has multiple problematic aspects. First,
it is questionable whether investments that are not attractive at already
negative real interest rates, but only get implemented when real interest
rates fall even further, will be productive. Second, these new and uncon-
ventional policies create uncertainty, as markets puzzle about the strategy
of winding down quantitative easing and about the effect of forward guid-
ance on investors’ beliefs. Third, as Jeremy Stein5 and others have pointed
out, reducing interest rates through unconventional monetary policy
4. See, for instance, IMF, “Will It Hurt? Macroeconomic Effects of Fiscal Con-
solidation,” chapter 3 of the IMF’s October 2010 “World Economic Outlook”;
Jaime Guajardo, Daniel Leigh, and Andrea Pescatori, “Expansionary Austerity:
New International Evidence,” IMF Working Paper WP/11/158 (2011); Christina
D. Romer and David H. Romer, “The Macroeconomic Effects of Tax Changes:
Estimates Based on a New Measure of Fiscal Shocks,” American Economic Review
100 (2010): 763–801; and Alan J. Auerbach and Yuriy Gorodnichenko, “Measuring
the Output Responses to Fiscal Policy,” American Economic Journal: Economic
Policy 4, no. 2 (2012): 1–27.
5. Jeremy C. Stein, “Overheating in Credit Markets: Origins, Measurement,
and Policy Responses,” speech at the “Restoring Household Financial Stability
after the Great Recession: Why Household Balance Sheets Matter” research sym-
posium sponsored by the Federal Reserve Bank of St. Louis, St. Louis, Missouri,
February 7, 2013.
Copyright © 2014 by the Board of Trustees of the Leland Stanford Junior University. All rights reserved.Low Equilibrium Real Rates, Financial Crisis, and Secular Stagnation 49
leads investors to increase risk-taking and leverage, thus raising the like-
lihood of bubbles. The argument that macroprudential policies can be
used to contain such financial excesses is a chimera—unconventional
monetary policy stimulates the economy precisely by increasing asset
values and the ability to borrow, which prudential regulation aims to
address. In addition, macroprudential policies rely on the ability of regu-
latory agencies to spot and curb bubbles—the same regulators who were
unable to discern that Lehman Brothers, Wachovia, Washington Mutual,
and Bear Stearns were undercapitalized even a week before they failed.
Regulatory approaches that do not require regulators to be able to out-
guess markets are preferable. These include sharp increases in capital and
liquidity requirements and the provision of swift and forceful resolution
authority. Finally, there are distributional concerns: policy measures that
drive down interest rates to inflate asset values benefit those who hold the
assets, which are disproportionately the wealthy.
The third and most promising policy option is to spur spending at
every possible level of the real interest rate. The most direct way to do
this is through fiscal policy action. Consider infrastructure investments,
which not only increase productivity and thereby raise GDP, but also
stimulate demand in an economy that is demand-constrained. As such,
fixing John F. Kennedy International Airport in an environment with a
construction unemployment rate in the double digits by issuing long-term
debt at very low interest rates should be highly attractive.
When the growth rate exceeds the interest rate, which will likely be
the case for a long time for short-term debt, the debt-to-GDP ratio will
decline if the government issues debt and rolls over the debt to cover
interest payments. Elsewhere, I have demonstrated6 that in the model of
the US economy used for the forecasting and analysis of monetary policy
at the Federal Reserve Board, a five-year fiscal impulse in the context
of a nominal short-term interest rate at the zero lower bound leads to a
lower debt-to-GDP ratio in twenty-five years. The model (1) assumes that
government expenses do not contribute to utility or productivity—they
are simply goods that are produced and contribute to GDP, but then are
thrown into the ocean; and (2) does not take into account public debt at
the local and state levels. Both are model simplifications that understate
the real-world effect of expansionary fiscal policy on GDP growth, so that
6. Lawrence Summers and David Reifschneider, ongoing analysis for speech
at National Association for Business Economics, 2014.
Copyright © 2014 by the Board of Trustees of the Leland Stanford Junior University. All rights reserved.50 Lawrence H. SummerS
the future debt-to-GDP ratio would likely be even lower than the model
predicts.
But public investment is just one way to boost demand. There are
ample opportunities to improve the efficiency of regulation in ways that
would stimulate demand, particularly in the energy sector. One example
in the United States would be to allow exports of fossil fuels. In general,
a concerted effort to promote competitiveness to increase net exports
would raise demand in a single nation without the need to lower interest
rates. However, such a strategy would not work for the world as a whole.
Finally, long-run supply-side fundamentals such as policy measures that
ensure the sustainability of entitlement programs, provide for tax reform,
and facilitate investments in labor force skills and innovation can contrib-
ute to confidence and thereby boost demand in the short term.
Conclusion
Economic developments over the past decade raise concerns about sec-
ular stagnation. In a sufficiently low inflation environment, it may be
impossible to attain real interest rates consistent with full employment.
Even if it is possible, monetary policy actions that keep short-term nomi-
nal interest rates near zero by reducing term and risk premiums raise the
likelihood of financial excesses and future crises. However, secular stagna-
tion is not inevitable. We can ensure both adequate economic growth and
financial stability with the right policy choice: a commitment to structural
increases in demand. Embracing this objective will require a sea change
in contemporary economic thinking.
Copyright © 2014 by the Board of Trustees of the Leland Stanford Junior University. All rights reserved.You can also read
