Working Paper Series
Congressional Budget Office
Washington, DC
The Fiscal Multiplier and Economic Policy Analysis in the United States
Charles J. Whalen
Macroeconomic Analysis Division
Congressional Budget Office
[email protected]
Felix Reichling
Macroeconomic Analysis Division
Congressional Budget Office
[email protected]
February 2015
Working Paper 2015-02
To enhance the transparency of the work of the Congressional Budget Office (CBO) and to encourage
external review of that work, CBO’s working paper series includes papers that provide technical
descriptions of official CBO analyses as well as papers that represent independent research by CBO
analysts. Papers in this series are available at http://go.usa/gov/ULE.
This paper is forthcoming in Contemporary Economic Policy. The authors received helpful comments and
suggestions from Douglas Elmendorf, Jeffrey Kling, Benjamin Page, Brad Humphreys (the journal editor),
and three anonymous referees.
Abstract
The recession from 2007 to 2009 sparked wide interest in the economic effects of fiscal policy.
That interest is reflected in an ongoing debate over the size of the fiscal multiplier. This working
paper addresses three questions: What models do economists use to estimate that multiplier? Why
do estimates of it vary widely? How can economists use those estimates to judiciously analyze
U.S. economic policy?
Contents
I. Introduction ................................................................................................................................. 1
II. What Models Do Economists Use to Estimate the Fiscal Multiplier? ....................................... 1
Macroeconometric Forecasting Models .................................................................................... 1
Times Series Models ................................................................................................................. 2
DSGE Models ............................................................................................................................ 3
III. Why Do Multiplier Estimates Vary Widely? .............................................................................. 4
Analytical and Measurement Issues .......................................................................................... 4
Fiscal Policy Details .................................................................................................................. 5
Economic Conditions ................................................................................................................ 6
Confidence ................................................................................................................................. 8
IV. How Can Multiplier Estimates Be Used to Analyze U.S. Economic Policy? ........................... 9
Methods and Policy Details ....................................................................................................... 9
Economic Conditions and Confidence .................................................................................... 10
V. Summary and Conclusion ......................................................................................................... 12
References ........................................................................................................................................ 13
Tables
Table 1. Ranges for U.S. Fiscal Multipliers ..................................................................................... 11
Table 2. The Effect of a $1 Increase in Aggregate Demand Over Eight Quarters ........................... 12
I. Introduction
The Great Recession, which began in December 2007 and ended in June 2009, sparked wide
interest in the economic effects of fiscal policy. The downturn initially provoked a flurry of papers
estimating how stimulus packages such as the American Recovery and Reinvestment Act of 2009
(ARRA) would affect output and employment.
1
Later, as many policymakers in the United States
and Europe sought to reduce government deficits, much attention shifted to the likely effects of
fiscal consolidation (increases in taxes and/or decreases in government spending or transfers).
2
In
addition, the entire period of recession and slow recovery has seen the release of numerous studies
examining how changes in fiscal policy affect economic outcomes.
3
The recent interest in how fiscal policy affects the economy is reflected in an ongoing debate over
the size of the fiscal multiplier, the change in a nation’s economic output generated by each dollar
of the budgetary cost of a change in fiscal policy. The multiplier must be estimated; it cannot be
observed.
4
Estimates of the fiscal multiplier vary widely, including values in excess of one and less
than zero.
5
What models do economists use to estimate the multiplier? Why do estimates of it vary widely?
And how can economists use those estimates to judiciously analyze U.S. economic policy? We
address the first two questions by reviewing the rapidly expanding body of academic literature and
address the third question by providing an overview of how the Congressional Budget Office
(CBO) uses multiplier estimates to analyze fiscal policy proposals and legislation.
II. What Models Do Economists Use to Estimate the Fiscal Multiplier?
Three types of models are often used to generate estimates of the fiscal multiplier—
macroeconometric forecasting models, time series models, and dynamic stochastic general
equilibrium (DSGE) models. Each type has strengths and limitations.
Macroeconometric Forecasting Models
Macroeconometric forecasting models, which underlie most of the forecasts offered to the clients
of economic consulting firms, are the basis for many estimates of multipliers. The details of those
models are based largely on historical relationships among aggregate economic variables and
informed by theories of how those variables are determined. Because macroeconometric
forecasting models emphasize the influence of the overall demand for goods and services, they
1
For example, see Fair (2010) and Romer and Bernstein (2009).
2
For example, see Gravelle and Hungerford (2013) and Bagaria et al. (2012).
3
For example, see Eggertsson (2009), Taylor (2011), and Fazzari et al. (2014).
4
Observing economic outcomes is not sufficient for determining the economic effects of a change in fiscal policy because isolating
the effects would require knowing what path the economy would have taken in the absence of the policy change.
5
For example, Van Brusselen (2009) reviewed literature on the U.S. economy and found estimates of government spending
multipliers between -3.8 and +3.8 and tax cut multipliers between -4.8 and +3.0. The suggestion that multipliers can be negative,
which means that cuts in spending and transfers would increase economic activity, is the argument behind some recent calls for
fiscal consolidation.
2
tend to estimate greater economic effects from policies that bolster demand than time series models
and DSGE models do.
6
The reliability of macroeconometric projections depends heavily on the validity of the specific
economic assumptions used. For example, because the models are grounded in observed historical
relationships, their estimates rely on the assumption that individuals will, on average, continue to
react to changes in fiscal policies in the same way that they reacted in the past. Consequently,
estimates projected by such models might be unreliable when policies or economic conditions
differ substantially from those of the past.
7
Times Series Models
Time series models offer an alternative to macroeconometric forecasting models. In their most
basic form, time series models, such as vector autoregression (VAR) models, summarize
correlations between economic variables—such as government spending and gross domestic
product (GDP)—over time.
8
Because time series models are grounded in historical data and
contain little economic theory, they can be particularly useful when there is reason to believe that
existing theories may be inaccurate or based on particularly unrealistic assumptions.
However, the lack of theoretical grounding makes it difficult to use time series models to assess the
direction of causation between policies and the economy. This is known as the “identification”
problem.
9
For example, while poor economic conditions can spur the government to enact policies
aimed at stimulating economic activity, a statistical correlation between the policies and economic
performance could be interpreted as indicating that policies caused the weak performance.
Two approaches are often used to identify economic causation as distinct from mere correlation.
One approach—called “structural vector autoregression” (SVAR)—relies on making assumptions
about the interaction of the economic variables of interest.
10
That approach is easy to implement
(because it does not require specification of many behavioral relationships or extensive data
gathering) and is useful when the statistical assumptions are correct. However, if the assumptions
are incorrect, then the approach may lead to less reliable multiplier estimates than the most basic
form of time series models. An alternative—“narrative”—approach supplements the time series
analysis of aggregate data with a review of historical evidence of other sorts, such as narrative
evidence from the legislative record. That alternative approach has been used most often to
6
For a more detailed discussion of macroeconometric models—and of time series models and DSGE models—see Chinn (2013).
7
See Parker (2011) and Auerbach et al. (2010) for a discussion of limitations that arise from the use of historical data to estimate
how output responds to new and untested fiscal policies.
8
The main difference between time series and macro models is that the latter impose a priori theory-based restrictions on the
relationship between model variables while time series models do not impose such restrictions, at least not in their purest form.
9
As in the case of macro models, the historical grounding of time series models can also make it difficult to assess what would
happen under economic conditions substantially different from those of the past. For further discussion of limitations and
complexities associated with using time series models, see Parker (2011).
10
For a discussion of the SVAR approach, see Blanchard and Perotti (2002).
3
estimate the economic impact of military buildups, events that are arguably unrelated to
macroeconomic conditions, but also to estimate the effects of tax changes.
11
DSGE Models
DSGE models are also used to estimate fiscal multipliers.
12
In DSGE models, people are assumed
to make decisions about how much to work, buy, and save on the basis of current and expected
future values of wage rates, interest rates, taxes, and government purchases, among other things.
As a result of those and other assumptions about individuals’ and businesses’ behavior, such
models offer a clear perspective on the causal relationships among economic variables.
13
A thorough grounding in economic theory allows DSGE models to avoid the difficulties of
interpretation that arise with purely statistical approaches to analyzing data. In addition, the explicit
assumptions about economic decision-making in DSGE models are less dependent on historical
data than in macroeconometric models.
14
Therefore, DSGE models can be particularly useful when
analyzing the effects of changes in fiscal policies that have not been observed previously.
DSGE models often include assumptions that seem at odds with important features of the real-
world economy.
15
For example, such models do not usually allow for underutilized resources in an
economy, such as involuntary unemployment or unused capital. In addition, people are generally
assumed to have full access to credit markets so that they can borrow to maintain their
consumption in the face of a temporary loss of income, and the Federal Reserve is often assumed
to respond to changes in fiscal policies, thereby excluding situations in which actions by the
Federal Reserve are constrained by a zero lower bound on nominal interest rates.
16
11
For examples of studies focusing on military buildups, see Ramey (2011a), and Ramey and Shapiro (1998), and for discussion of
the challenges associated with focusing on military buildups (namely, factors that could raise or dampen multipliers in such
periods), see Ramey (2011b, p. 677). Some articles in this literature (such as Owyang et al. 2013) do not adjust results for tax
increases that have often accompanied military buildups, but others make such an adjustment to generate results more applicable
to the case of deficit-financed government spending (including Ramey 2011a). For examples of studies that use the narrative
approach to estimate the effects of tax changes on economic activity, see Romer and Romer (2010), and Favero and Giavazzi
(2012).
12
DSGE models are “dynamic” because they focus on how an economy evolves over time, “stochastic” because they take into
account that the economy is affected by random shocks (owing to technological change, for example), and “general equilibrium”
because they assume that people make decisions in response to prices in the economy (such as wages and rates of return on
saving) and that prices change in response to those decisions.
13
DSGE models differ from traditional macro models in that they include micro-founded elements describing the optimal behavior
of economic agents.
14
DSGE models are generally calibrated so that macroeconomic variables, such as the total amount of labor supplied and the size of
the capital stock, match the amounts in the U.S. economy, or they are estimated using aggregate data to determine some key
parameters. See Fernández-Villaverde and Rubio-Ramírez (2006) for a detailed discussion of how DSGE models are estimated.
See Coenen et al. (2012) for a comparison of significant model features and parameters of several DSGE models used by
policymaking institutions in Canada, Europe, and the United States.
15
For example, see Parker (2011) and Fair (2012), who criticize several modeling choices made in many DSGE models. In addition,
Leeper et al. (2011) observe that a tight range for estimates of the multiplier is imposed by the assumptions and choices made by
researchers when using DSGE models. See also Chari et al. (2009), who argue that DSGE models rely on so many improvised
modeling assumptions that their conclusions are unavoidably ambiguous for policy analysis.
16
DSGE models also are typically built on the assumptions that people have full information about the current economy and future
economic developments and that they logically base their current decisions on a full lifetime plan. In extreme form, those
assumptions imply that people anticipate that increases in government spending or decreases in taxes will eventually lead to lower
4
However, additional research has relaxed many of the standard assumptions of DSGE models in an
effort to align those models more closely with important aspects of the economy. For example,
some models incorporate so-called “hand-to-mouth” consumers. Research on consumer behavior
finds that some households’ spending tends to vary one-for-one with income, perhaps in part
because those households have only small savings and face borrowing constraints and therefore
cannot maintain their desired level of consumption when their income falls, or because some
households follow simple behavior rules rather than trying to continuously determine their optimal
spending and saving. Recent multiplier estimates in models with such consumers are as much as
50 percent larger than estimates generated using standard DSGE models.
17
III. Why Do Multiplier Estimates Vary Widely?
The variation in estimates of the fiscal multiplier cannot be explained by economists’ use of
different types of models. Each type described above can generate a broad range of multiplier
estimates. For example, Reichling and Whalen (2012) find estimates for the United States, as
measured (on a cumulative basis) after eight quarters, ranging from 0.75 to 2.25 for
macroeconometric forecasting models, from 0.3 to 3.5 for time series models, and from 0.5 to 2.25
for DSGE models. To better understand the variation in multiplier estimates, one must consider
analytical and measurement issues, fiscal policy details, economic conditions, and how fiscal
policy can affect people’s confidence in the future of economic activity.
Analytical and Measurement Issues
Some variation in multiplier estimates is the result of differences in methodological choices, data
sets, and (as illustrated in the previous section) underlying behavioral assumptions. For example,
estimates can vary with the method used to address the identification problem and with the
approach used to measure a given variable. The significance of such methodological choices is
highlighted by Riera-Crichton et al. (2012), who analyze tax increases in 14 industrial countries
and find a tax multiplier of 1.32 (after three quarters) using the SVAR approach and a multiplier of
2.76 using the narrative approach (owing to policy anticipation); they also produce different
estimates when using alternatives to their preferred method of measuring tax policy instruments.
18
In addition, some variation in multiplier estimates is because analyses can differ with respect to the
period over which the multiplier is measured. Reported multipliers are sometimes “peak”
multipliers, which represent the largest effect on output in any one quarter after a policy change,
but others are “instantaneous” or “impact” multipliers (the former looks at the effect immediately
following a policy change and the latter allows for a lagged response). Still others are “cumulative”
multipliers, which represent the cumulative effect on output of a policy change over a given period.
Measuring the cumulative effect is often important because the effects on output can reverse
spending or higher taxes and that they raise their current saving in an attempt to offset that expected future burden. Therefore, in
such models, cash transfer payments and many sorts of reductions in taxes usually have little or no effect on current spending.
17
See, for example, Coenen et al. (2012) and Galí et al. (2007). Other researchers relax different assumptions; for example,
Fernández-Villaverde (2010) develops a model that incorporates financial frictions, and Leeper et al. (2009) study the economic
effects of government investment.
18
On the role of data sets, see, for example, Ramey (2011a), who calculates different estimates depending on whether World War II
is excluded from samples.
5
direction over time (as discussed further in Section IV). However, efforts to compare different
estimates are often complicated by the fact that studies sometimes “fail to specify the exact
multiplier concept used” (Auerbach et al. 2010, p. 114).
19
An especially challenging measurement issue stems from the possibility that households and
businesses act in anticipation of government spending and tax changes. The extent to which people
have “fiscal foresight” can have a significant effect on multiplier estimates (Leeper et al. 2012).
But that foresight is not easy to measure. Part of the difficulty is that time elapses between the
point at which there is recognition of the need for policy action and the point at which tax or
spending programs take effect—sometimes it is a short period, but other times it is a period of four
or more quarters. Another variable dimension is what Leeper et al. (2012, p. 130) call “foresight
intensity”: how confident people are about pending changes. Moreover, the effects of fiscal
foresight are not unambiguous; for example, the ways in which people act in anticipation of a fiscal
policy change can depend on policy details and economic conditions.
Fiscal Policy Details
Looking beyond analytical and measurement issues, estimates of the fiscal multiplier can vary
because the size of the multiplier depends on fiscal policy details—including the nature, duration,
and timing of policy changes. Also important to multiplier estimates are people’s expectations
about the fiscal policy details yet to come: the future path of government spending and revenue.
As in the case of Van Brusselen (2009), the economics literature often makes a distinction between
tax and government spending multipliers and frequently finds that the spending multipliers are
larger than tax multipliers.
20
To understand why, consider two changes to fiscal policy—an
increase in government spending and a cut in taxes, each with a budgetary cost of a dollar. The
increase in government spending immediately contributes a dollar to aggregate demand, but the tax
cut (or, alternatively, an increase in transfer payments) could contribute less than a dollar because
it can be spent or saved (the marginal propensity to consume can be less than one). It is this
difference that causes some analysts to distinguish between government spending and tax
multipliers.
Theory and some empirical evidence also suggest that finer multiplier distinctions may be
warranted. This includes separate multipliers for public investment and public consumption. For
example, Auerbach and Gorodnichenko (2012a) find that multipliers for infrastructure spending
and other types of U.S. public investment are larger than multipliers for public consumption.
21
In addition, multipliers can vary across different fiscal policy provisions because those provisions
can affect people with different characteristics, resulting in different responses. For example, tax
cuts are likely to boost purchases more for lower-income households than for higher-income
households.
22
That difference arises, at least in part, because lower-income households typically
19
There are also other types of multipliers, such as present-value multipliers; see Leeper et al. (2010).
20
For example, see also Chahrour et al. (2012) and Leeper et al. (2011).
21
Similar results are found in recent studies that focus on public investment outside of the United States. See Gonzalez-Garcia et al.
(2013) and Bruckner and Taladhar (2010). However, an earlier study by Perotti (2004) found no evidence of larger multipliers for
public investment (relative to public consumption) in five countries, including the United States.
22
For additional examples, see Zandi (2011).
6
consume a higher fraction of their income and because they are less able to borrow money to
finance their desired consumption.
The duration and timing of fiscal policy actions can also affect the size of the multiplier. For
example, Coenen et al. (2012) find that fiscal stimulus has a larger effect when it is of “moderate
persistence”—namely, two or three years in duration. In particular, they suggest that a policy
change of moderate duration boosts consumer spending more than a one-time action, and that some
expansionary effects of an even more persistent policy initiative can be offset by concerns about
rising future taxes. In addition, Christiano et al. (2011) show that multipliers are larger when all—
rather than just a portion—of the government spending in a fiscal stimulus is timed to coincide
with a zero-bound constraint on interest rates.
23
The multiplier can also be affected by people’s expectations about the future path of government
spending and revenue. For example, consider the case of a stimulus package that increases
government spending. The way people expect that package to be financed in the future—by means
of changes in taxes (on labor, capital, or consumption), transfers, government spending, or a
certain combination of those alternatives—represents an important set of fiscal policy details; and
different assumptions about such details, including the speed of policy adjustment, will result in
different multiplier estimates (Leeper et al. 2010).
Economic Conditions
Estimates of the fiscal multiplier also vary with economic conditions, including the state of the
business cycle, the response of monetary policy to fiscal policy changes, and the condition of the
financial system.
24
Some recent economic research finds that the state of the business cycle affects the size of the
fiscal multiplier. For example, Auerbach and Gorodnichenko (2012a) extend an SVAR model to
allow for responses differentiated across recessions and expansions and use that regime-switching
model to estimate a peak multiplier for government spending in the United States of 2.5 in
recessions and 0.6 in expansions. That finding can be explained as follows: When the economy’s
labor and capital resources are close to being fully utilized, some of the effect of a fiscal stimulus
could be offset by a reduction in private-sector spending that would have occurred in the absence
of the stimulus (owing to a bidding up of the price of the economy’s resources); however, when
there are more unused resources, any such reduction is likely to be smaller and a fiscal stimulus is
more likely to be magnified by additional spending by the private sector.
25
23
See also Coenen et al. (2012). In contrast, Cogan et al. (2010) assume that only a small fraction of an increase in government
spending would occur when the short-term interest rate is at the zero lower bound and that the rest would occur after the short-
term interest rate begins to rise; as a result, they estimate smaller multipliers.
24
In addition to the economic conditions discussed in this article, which focuses on the U.S. economy, a broader discussion would
include structural characteristics such as an economy’s size, its openness to trade, and the nature of its exchange rate regime. For a
discussion of how such characteristics affect fiscal multipliers, see Chinn (2013) and Batini et al. (2014). See also Ilzetzki et al.
(2013), who classify the United States as a relatively large, closed economy according to the size of its internal market and
openness to trade (they designate an economy as “closed” if foreign trade is less than 60 percent of GDP) and find that fiscal
multipliers are larger in large, closed economies than in small, open economies.
25
Some economists have suggested that interest rates must be constrained by the zero lower bound for fiscal multipliers to be larger
in recessions than in expansions. However, Michaillat (2012) develops a labor-market search model with job rationing to show
that there can be those larger multipliers whenever the labor market is depressed.
7
In a follow-up paper, Auerbach and Gorodnichenko (2012b) confirm their general results (larger
multipliers in recessions than in expansions) by extending the analysis to a large number of
countries that are members of the Organisation for Economic Co-operation and Development
(OECD) and by using an approach known as expectations-augmented VAR. Other studies that use
different types of VAR models also find fiscal multipliers are larger in times of economic slack
than in robust expansions, including Fazzari et al. (2014); and Baum et al. (2012).
In contrast, Owyang et al. (2013) and Ramey and Zubairy (2014) do not find larger multipliers in
times of slack when using a narrative approach to examine the effect of changes in government
spending in the wake of news about military events. That contrast underscores a point made above:
analysts’ choices about methods, measurement, and data can play a decisive role in generating
multiplier estimates. As Ramey and Zubairy (2014, p. 3) write, “Most of the differences in
conclusions between our work and that of Auerbach and Gorodnichenko lie in ... the construction
of impulse response functions on which the multipliers are based. In contrast to linear models,
where the calculation of impulse response functions is a straightforward undertaking, constructing
impulse response functions in nonlinear models is fraught with complications.”
26
Some economists—including Hall (2009), Christiano et al. (2011), Davig and Leeper (2011), and
Coenen et al. (2012)—also find that the size of the fiscal multiplier depends on the response of
monetary policy to fiscal policy changes.
27
Under normal economic conditions, a fiscal stimulus
would increase inflationary pressure and the Federal Reserve would increase the federal funds rate
as a countermeasure. However, there can be times when monetary policy is constrained by the zero
lower bound and fiscal policy changes would not be offset by changes in the funds rate.
28
Indeed,
the Federal Reserve kept short-term interest rates near zero as federal fiscal policymakers adopted
stimulus measures during and after the Great Recession. For example, Hall (2009) finds that the
government spending multiplier (after four quarters) rises from just below 1.0 to 1.7 when
monetary policy is passive because the federal funds rate is at its lower bound of zero.
29
The size of the multiplier can also be affected by the condition of the financial system. For
example, using a panel of OECD countries, Corsetti et al. (2012) find larger fiscal multipliers
during times of financial crisis.
30
They suggest that is because spending by the private sector is
more likely to be constrained by a lack of access to credit during such times.
26
An even more recent paper by Riera-Crichton et al. (2014) revisits the question of whether multipliers are larger in recessions
than in expansions and brings in a new dimension: whether government spending is going up or down in bad times. Using a
sample of 29 OECD countries, they conclude that “the ‘true’ long-run multiplier for bad times (and government spending going
up) turns out to be 2.3 compared to 1.3 if we just distinguish between recession and expansion. In extreme recessions, the long-run
multiplier reaches 3.1.”
27
As in the case of whether multipliers vary with the state of the business cycle, Ramey and Zubairy (2014) look at episodes of
increased U.S. military spending and find little evidence that the fiscal multiplier is higher at the zero lower bound.
28
For a discussion of the economic mechanisms behind such results, see Woodford (2011).
29
Erceg and Lindé (2012) also find larger multipliers at the zero lower bound, provided that the size of fiscal stimulus is small. As
the size of the stimulus gets larger (Erceg and Lindé use a threshold value of 1.2 percent of GDP, for example), they find the
multiplier falls because the period during which the economy is constrained by the zero lower bound is shortened.
30
Similarly, Afonzo et al. (2011) look at the United States, United Kingdom, Germany, and Italy between 1981 and 2009 and find
higher multipliers during periods of financial stress.
8
Confidence
Multiplier estimates can also vary as a consequence of how changes in fiscal policy affect people’s
confidence about the future of the economy. For example, using a VAR approach, Bachmann and
Sims (2012) find that increases in government spending during downturns boost consumer
confidence, which, in turn, increases multipliers. Moreover, they find that such increases in
spending, often geared toward public investment, boost long-term productivity. As a result, they
argue the increase in confidence reflects that productivity boost, rather than pure sentiment.
Other research suggests that changes in fiscal policy can have a different effect on confidence.
Motivated by a growing body of literature that suggests uncertainty negatively affects economic
activity, Alloza (2014) uses both SVAR and narrative approaches to examine whether fiscal policy
is more effective in tranquil times (“low uncertainty”) than in more uncertain times (“high
uncertainty”). He finds smaller multipliers in the more uncertain periods and identifies households’
confidence as the key variable in explaining that result: An increase in government spending in
such times confirms people’s views about economic weakness and results in a decline in consumer
spending.
31
Some research even suggests that changes in fiscal policy can affect people’s confidence in a way
that produces negative multipliers. For example, Corsetti et al. (2013) outline what they call
“extreme cases” in which an increase in government spending threatens government solvency. In
such cases, they find that the multiplier could be negative because the increased risk of a sovereign
default reduces private-sector demand. They also suggest that this process could work in the other
direction in the case of fiscal consolidation: If such consolidation reduces the risk of a sovereign
default, then that policy change could bolster economic activity.
32
However, much of the recent literature looks skeptically at the notion of expansionary fiscal
consolidations. For example, although Alesina and Ardagna (2010) provide some evidence from
OECD countries that fiscal consolidation has sometimes resulted in an output expansion, two
subsequent reports by analysts at the International Monetary Fund (Guajardo et al. 2011;
International Monetary Fund 2010) find that consolidation more typically reduces output and that
measures used in the earlier study bias the results toward overstating expansionary effects.
Moreover, Perotti (2011) looks closely at four case studies involving expansionary fiscal
consolidations (Denmark, 1983–1986; Ireland, 1987–1989; Finland, 1992–1998; Sweden, 1993–
1998) and concludes that three of the expansions were driven by currency depreciation and an
export boom, not by a boost in confidence, while the fourth case—the Danish expansion—was
short-lived (and was followed by a loss of competitiveness and a six-year slump).
31
Of course, one could provide a plausible interpretation had Alloza found the opposite results; for example, in that case one could
argue that fiscal stimulus reduces uncertainty and boosts confidence by limiting the damaging hysteresis effects that can occur in a
downturn. For a discussion of hysteresis and fiscal policy, see DeLong and Summers (2012).
32
For a similar analysis, see Müller (2014). See also Alesina and Ardagna (2010), who suggest that fiscal adjustment could be
expansionary if people believe that fiscal tightening represents a change in regime that eliminates the need for larger, more
disruptive adjustments in the future. To be sure, effects on confidence—or on similar psychological dimensions, such as
uncertainty (Blanchard 1990) and expectations (Bertola and Drazen 1993; Sutherland 1997)—are not the only possible
explanations for a negative multiplier. For example, Corsetti et al. (2013) highlight the effect of changes in fiscal policies on
private-sector funding costs (interest rates and credit risk premiums). However, much academic and policy attention has focused
on the “confidence channel,” especially in the case of fiscal consolidations. See, for example, Perotti (2011) and Chinn (2013).
9
IV. How Can Multiplier Estimates Be Used to Analyze
U.S. Economic Policy?
Despite economists’ disagreement about the size (and even the direction) of the multiplier,
informed policymaking requires sound estimates of the economic effects of proposed fiscal policy
actions. CBO, a nonpartisan agency within the legislative branch of the U.S. government, regularly
confronts that challenge. CBO’s mission is to provide Congress with objective, independent, and
timely information and analyses. Consistent with its mandate, CBO makes no recommendations;
but it is often asked to analyze the potential role and efficacy of fiscal policy options in influencing
output and employment.
Methods and Policy Details
CBO uses estimates of the fiscal multiplier to analyze the impact that fiscal policy action has on
economic output by means of its influence on the overall demand for goods and services.
33
Although much of the research on the size of fiscal multipliers focuses on multipliers associated
with increased government spending and large-scale tax cuts, CBO is asked to analyze the
economic impacts of a wide array of possible fiscal policy changes. To bridge that gap, the agency
uses an estimation method that can accommodate different policy details.
CBO’s fiscal multiplier is the product of two separate effects. In particular, CBO decomposes the
multiplier into a direct effect (the effect of a dollar of transfers or spending on the demand for
goods and services) and an indirect effect (the effect on output that arises when the direct effects
propagate throughout the economy). Using that approach, the effect on output of a dollar change in
fiscal policy can be written as follows:
($1 Change in Budgetary Cost)
*
[Fiscal Multiplier] =
($1 Change in Budgetary Cost)
*
[(Direct Effect on Demand)
*
(Indirect Effect on
Demand)] = Change in Output.
The direct effect of a change in fiscal policy depends on the policy details. In the case of a dollar
increase (decrease) in government purchases, the direct effect is 1 because demand increases
(decreases) by a dollar. In the case of a dollar increase (decrease) in taxes or transfer payments, the
direct effect can vary considerably across fiscal policy provisions (largely because they affect
people with different characteristics and may be seen as more or less persistent, resulting in
different responses, as discussed above). For example, CBO’s analysis of ARRA relies on
estimates of direct effects for eight major tax and transfer provisions of that legislation
(Congressional Budget Office 2014a).
33
CBO analyzes the effects of changes in federal fiscal policies on the economy in the short term and the longer term. In the short
term, changes in fiscal policies affect economic output primarily by influencing the demand for goods and services. In the longer
term, a period in which CBO assumes that actual output is close to potential output, changes in fiscal policies affect output
primarily by altering the incentives for individuals and businesses to work, save, and invest. If the labor market is sufficiently
tight, changes in incentives to work also can affect output and employment in the short term, but this effect has usually been small
in CBO’s analyses.
10
The indirect effect of a change in policy offsets or enhances the direct effect. For example, the
direct effects of lower taxes or higher government spending are magnified when stronger demand
for goods and services prompts businesses to increase investment and hire more workers than they
otherwise would. In the other direction, the direct effects are muted if—for example—higher
government borrowing caused by tax cuts or spending increases leads to higher interest rates that
discourage (“crowd out”) spending on investment and durable goods such as cars (because higher
interest rates raise the cost of borrowing by households and businesses).
34
Thus, the indirect effect
represents the total change in output per dollar of direct effect on demand.
Given the wide range of multiplier estimates and the uncertainty about the economic relationships
that underlie their estimation, CBO uses a range of estimates in its analyses (often in conjunction
with a central estimate). That range encompasses a broad spectrum of economists’ views about the
underlying economic relationships and is informed by results generated by macroeconometric
models, time series models, and DSGE models. In particular, the upper portions of CBO’s ranges
are based mainly on macro models developed by Macroeconomic Advisers and IHS Global
Insight; the lower ends are based mainly on the literature using time series models; and CBO relies
on DSGE models primarily to help understand the economic and behavioral mechanisms that
underlie estimates in the empirical literature (such as anticipation of policy action, as discussed in
Section III) and to gauge how changes in business and consumer behavior may affect multipliers.
CBO’s analysis of ARRA illustrates the results that can be produced by combining estimates of
direct and indirect effects (Congressional Budget Office 2014a). In particular, CBO’s ARRA report
presents a range of fiscal-multiplier estimates for each major provision of that legislation (see
Table 1). In that report, purchases of goods and services by the federal government had the highest
estimated fiscal multiplier (estimates ranged from 0.5 to 2.5) and a set of corporate tax provisions
had the lowest estimated multiplier (estimates ranged from 0 to 0.4).
35
Economic Conditions and Confidence
Reflecting CBO’s assessment of the literature surveyed in Section III, the magnitudes of the
multiplier that the agency uses vary with economic conditions. For example, when output is well
below its potential and the Federal Reserve’s response to changes in fiscal policies is likely to be
limited (such as in recent years when unemployment was elevated, inflation was low, and the
Federal Reserve’s ability to reduce interest rates was constrained because those rates were already
near zero), CBO estimates that multipliers are larger than when output is close to or above its
potential and the Federal Reserve responds more fully to counteract the effects of changes in fiscal
policies.
34
Private-sector spending could also be “crowded out” through other channels besides an increase in interest rates. For example,
activities spurred by stimulative fiscal policies could reduce production elsewhere in the economy if they used scarce materials or
workers with specific skills and thereby created bottlenecks that hindered other production. As with crowding out caused by rising
short-term interest rates, crowding out caused by production bottlenecks has probably been much smaller during the most recent
U.S. recession and slow recovery (because of high unemployment and a large amount of unused capital) than it might be during
other periods. Another channel for crowding out is that some people will respond to a fiscal stimulus by cutting back their
spending in anticipation of higher taxes in the future (as mentioned in the previous section).
35
For other CBO analyses of the economic effects of fiscal policies, see, for example, Congressional Budget Office (2012; 2014b).
Also, for discussions of the economic effects of federal investment, see Congressional Budget Office (2013, pp. 4–5).
11
Table 1.
Ranges for U.S. Fiscal Multipliers
Source: Congressional Budget Office (CBO).
Note: The estimates above were produced for CBO's analysis of the American Recovery and Reinvestment Act
of 2009.
CBO recently published a range of estimates for changes in output resulting from a dollar increase
(decrease) in aggregate demand (see Table 2).
36
In the case of tax cuts and increases in transfer
payments, those ranges represent the “indirect effect” portion of the fiscal multiplier; and in the
case of a change in government purchases, the ranges represent the entire fiscal multiplier (because
the direct effect equals 1 for such purchases). When output is well below its potential and Federal
Reserve responses are likely to be limited, CBO estimates that a dollar increase (decrease) in
demand will increase (decrease) output over four quarters—beginning in the first quarter in which
a direct effect occurs—and that the cumulative effect on GDP over that period will range from 0.5
(which means the direct effect is muted) to 2.5 (which means the direct effect is magnified).
In contrast, when output is close to its potential and the Federal Reserve responds more typically to
changes in fiscal policies, CBO estimates that a dollar increase (decrease) in demand will have
effects over eight quarters. Over the first four quarters, CBO estimates that the cumulative effect
on output will be similar to when output is well below potential, but the cumulative effect on GDP
over eight quarters ranges from 0.2 to 0.8 (as shown in Table 2). Those values are smaller than
when output is well below potential because the economic impact of changes in interest rates
grows over time and output in quarters five through eight moves in in the opposite direction of its
initial path.
CBO’s analyses are also informed by the literature on multipliers and confidence (reviewed in
Section III). In particular, although CBO has estimated that the multiplier is not negative currently,
the agency recognizes there can be situations that would cause multipliers to be negative. Indeed,
that issue was explored in a report CBO published in 2010 (Congressional Budget Office 2010).
36
CBO expects that the economic effects of changes in fiscal policies are roughly symmetric, meaning that under similar economic
conditions the size of the fiscal multiplier is the same for stimulative polices (such as increases in government spending or
decreases in taxes) as for contractionary policies (such as decreases in government spending or increases in taxes).
0.5
2.5
Transfer Payments to State and Local Governments for Infrastructure
0.4
2.2
0.4
1.8
Transfer Payments to Individuals
0.4
2.1
One-Time Payments to Retirees
0.2
1.0
Two-Year Tax Cuts for Lower- and Middle-Income People
0.3
1.5
One-Year Tax Cut for Higher-Income People
0.1
0.6
Extension of First-Time Homebuyer Credit
0.2
0.8
Corporate Tax Provisions Primarily Affecting Cash Flow
0
0.4
Purchases of Goods and Services by the Federal Government
Transfer Payments to State and Local Governments for Other Purposes
Type of Activity
Low Estimate
High Estimate
Estimated Multipliers
12
Table 2.
The Effect of a $1 Increase in Aggregate Demand Over Eight Quarters
Source: Congressional Budget Office (CBO).
Notes: CBO published these estimates on October 11, 2012, as a supplement to Felix Reichling and Charles
Whalen, Assessing the Short-Term Effects on Output of Changes in Federal Fiscal Policies, CBO Working Paper
2012-08 (May 2012). In that data release, “When Output Is Well Below Potential . . .” is described as “When
Short-Term Interest Rates Are Close to Zero” and “When Output Is Close to Potential . . .” is described as “When
Short-Term Interest Rates Are Not Close to Zero.” See www.cbo.gov/publication/43278.
There are no effects after eight quarters.
V. Summary and Conclusion
The Great Recession sparked wide interest in the economic effects of fiscal policy. That recent
interest is reflected in an ongoing debate over the size of the fiscal multiplier. This survey article
addressed three questions: What models do economists use to estimate that multiplier? Why do
estimates of it vary widely? And how can economists use those estimates to judiciously analyze
U.S. economic policy?
Three kinds of models are often used to generate estimates of the fiscal multiplier:
macroeconometric forecasting models, time series models, and DSGE models. Each has strengths
and limitations. For example, estimates generated by macroeconometric models benefit from being
grounded in historical data and economic theory, but they also might be unreliable when policies or
economic conditions differ substantially from the past.
A key finding above is that the variation in multiplier estimates cannot be explained entirely by
economists’ use of different types of models. Estimates vary widely primarily because analysts use
a variety of estimation methods and because estimates can depend on the details of fiscal policy,
the nature of economic conditions, and how fiscal policy affects confidence in economic activity.
Consistent with that finding, CBO varies its multiplier ranges for different economic conditions.
CBO’s approach to estimating the fiscal multiplier involves first distinguishing between the direct
and indirect effects of fiscal policy changes and then combining estimates of both effects. The
result is set of ranges informed by a variety of models and the rapidly expanding body of literature.
When Output Is Well Below
Low Estimate
High Estimate
Responses Are Typical
When Output Is Close to
Potential And Federal Reserve
Potential And Federal Reserve
Low Estimate
High Estimate
Responses Are Limited
13
References
Afonzo, Antonio, Jaromir Baxa, and Michal Slavik. 2011. “Fiscal Developments and Financial
Stress: A Threshold VAR Analysis.” European Central Bank Working Paper, No. 1319.
Alesina, Alberto and Silvia Ardagna. 2010. “Large Changes in Fiscal Policy: Taxes Versus
Spending,” in Tax Policy and the Economy, Jeffrey R. Brown, ed., vol. 24 (Chicago: University of
Chicago Press), pp. 35–68.
Alloza, Mario. 2014. “Is Fiscal Policy More Effective in Uncertain Times or During Recessions?”
University College London, http://tinyurl.com/ootkv8z.
Auerbach, Alan J., William G. Gale, and Benjamin H. Harris. 2010. “Activist Fiscal Policy.”
Journal of Economic Perspectives, 24 (4): 141–164.
Auerbach, Alan J. and Yuriy Gorodnichenko. 2012a. “Measuring the Output Responses to Fiscal
Policy.” American Economic Journal: Economic Policy, 4 (2): 1–27.
Auerbach, Alan J. and Yuriy Gorodnichenko. 2012b. “Fiscal Multipliers in Recession and
Expansion.” Working Paper (January 2012), http://tinyurl.com/lw5tlax.
Bachmann, Rudiger and Eric R. Sims. 2012. “Confidence and the Transmission of Government
Spending Shocks.” Journal of Monetary Economics, 59 (3): 235–249.
Bagaria, Nitika, Dawn Holland, and John Van Reenen. 2012. “Fiscal Consolidation During a
Depression.” Centre for Economic Performance Special Paper.
Batini, Nicoletta, Luc Eyraud, and Anke Weber. 2014. “A Simple Method to Compute Fiscal
Multipliers.” International Monetary Fund Working Paper, No. 14–93.
Baum, Anja, Marcos Poplawski-Ribeiro, and Anke Weber. 2012. “Fiscal Multipliers and the State
of the Economy.” International Monetary Fund Working Paper, No. 12–286.
Bertola, Giuseppe and Allan Drazen. 1993. “Trigger Points and Budget Cuts: Explaining the
Effects of Fiscal Austerity.” American Economic Review, 83 (1): 11–26.
Blanchard, Olivier Jean. 1990. “Comment on Can Severe Fiscal Contractions Be Expansionary?”
in NBER Macroeconomics Annual 1990, Oliver J. Blanchard and Stanley Fischer, eds., vol. 5
(Cambridge, Massachusetts: MIT Press), pp. 111–117.
Blanchard, Olivier and Roberto Perotti. 2002. “An Empirical Characterization of the Dynamic
Effects of Changes in Government Spending and Taxes on Output.” The Quarterly Journal of
Economics, 117 (4): 1329–1368.
Bruckner, Markus and Anita Tuladhar. 2010. “Public Investment as a Fiscal Stimulus: Evidence
From Japan’s Regional Spending During the 1990s.” International Monetary Fund Working Paper,
No. 10–110.
14
Chahrour, Ryan, Stephanie Schmitt-Grohé, and Martín Uribe. 2012. “A Model-Based Evaluation
of the Debate on the Size of the Tax Multiplier.” American Economic Journal: Economic Policy, 4
(2): 28–45.
Chari, Varadarajan V., Patrick J. Kehoe, and Ellen R. McGrattan. 2009. “New Keynesian Models:
Not Yet Useful for Policy Analysis.” American Economic Journal: Macroeconomics, 1 (1): 242–
266.
Chinn, Menzie. 2013. “Fiscal Multipliers.” The New Palgrave Dictionary of Economics, Steven N.
Durlauf and Lawrence E. Blume, eds. Palgrave Macmillan.
Christiano, Lawrence, Martin Eichenbaum, and Sergio Rebelo. 2011. “When Is the Government
Spending Multiplier Large?” Journal of Political Economy, 119 (1): 78–121.
Coenen, Günter, Christopher Erceg, Charles Freedman, Davide Furceri, Michael Kumhof, René
Lalonde, Douglas Laxton, Jesper Lindé, Annabelle Mourougane, Dirk Muir, Susanna Mursula,
Carlos de Resende, John Roberts, Werner Roeger, Stephen Snudden, Mathias Trabandt, and Jan
in’t Veld. 2012. “Effects of Fiscal Stimulus in Structural Models.” American Economic Journal:
Macroeconomics, 4 (1): 22–68.
Cogan, John F., Tobias Cwik, John B. Taylor, and Volker Wieland. 2010. “New Keynesian Versus
Old Keynesian Government Spending Multipliers.” Journal of Economic Dynamics and Control,
34 (3): 281–295.
Congressional Budget Office. 2010. Federal Debt and the Risk of a Fiscal Crisis.
Congressional Budget Office. 2012. Economic Effects of Policies Contributing to Fiscal
Tightening in 2013.
Congressional Budget Office. 2013. Federal Investment.
Congressional Budget Office. 2014a. Estimated Impact of the American Recovery and
Reinvestment Act on Employment and Economic Output in 2013.
Congressional Budget Office. 2014b. The Economic Effects of the President’s 2015 Budget.
Corsetti, Giancarlo, Andre Meier, and Gernot J. Muller. 2012. “What Determines Government
Spending Multipliers?” International Monetary Fund Working Paper, No. 12–150.
Corsetti, Giancarlo, Keith Kuester, Andre Meier, and Gernot J. Muller. 2013. “Sovereign Risk,
Fiscal Policy, and Macroeconomic Stability.” The Economic Journal, 123 (February): F99–F132.
Davig, Troy, and Eric M. Leeper. 2011. “Monetary-Fiscal Policy Interactions and Fiscal Stimulus.”
European Economic Review, 55 (2): 211–227.
Delong, J. Bradford and Lawrence H. Summers. 2012. “Fiscal Policy in a Depressed Economy.”
Brookings Papers on Economic Activity, 44 (1): 233-274.
15
Eggertsson, Gauti B. 2009. “What Fiscal Policy is Effective at Zero Interest Rates?” Federal
Reserve Bank of New York Staff Report, No. 402.
Erceg, Christopher and Jesper Lindé. 2012. “Is There a Fiscal Free Lunch in a Liquidity Trap?”
Board of Governors of the Federal Reserve System International Discussion Papers, No. 1003r.
Fair, Ray C. 2010. “Estimated Macroeconomic Effects of the U.S. Stimulus Bill.” Contemporary
Economic Policy, 28 (4): 439–452.
Fair, Ray C. 2012. “Has Macro Progressed?” Journal of Macroeconomics, 34 (1): 2–10.
Favero, Carlo and Francesco Giavazzi. 2012. “Measuring Tax Multipliers: The Narrative Method
in Fiscal VARs.” American Economic Journal: Economic Policy, 4 (2): 69–94.
Fazzari, Steven M., James Morley, and Irina B. Panovska. 2014. “State-Dependent Effects of
Fiscal Policy.” University of New South Wales Business School Research Paper, No. 2012, 27C.
Fernández-Villaverde, Jesús. 2010. “Fiscal Policy in a Model With Financial Frictions.” American
Economic Review: Papers & Proceedings, 100 (2): 35–40.
Fernández-Villaverde, Jesús and Juan F. Rubio-Ramírez. 2006. “The Research Agenda: Estimating
DSGE Models.” Economic Dynamics Newsletter, 8 (1), http://tinyurl.com/lhorxp4.
Galí, Jordi, J. David López-Salido, and Javier Vallés. 2007. “Understanding the Effects of
Government Spending on Consumption.” Journal of the European Economic Association, 5 (1):
227–270.
Gonzalez-Garcia, Jesus, Antonio Lemus, and Mico Mrkaic. 2013. “Fiscal Multipliers in the
ECCU.” International Monetary Fund Working Paper, No. 13–117.
Gravelle, Jane G. and Thomas L. Hungerford. 2013. Can Contractionary Fiscal Policy Be
Expansionary? Congressional Research Service, Report for Congress, No. R418949.
Guajardo, Jamie, Daniel Leigh, and Andrea Pescatori. 2011. “Expansionary Austerity: New
International Evidence.” International Monetary Fund Working Paper, No. 11
–158.
Hall, Robert E. 2009. “By How Much Does GDP Rise if the Government Buys More Output?”
Brookings Papers on Economic Activity 2:2009, pp. 183–250.
Ilzetzki, Ethan, Enrique G. Mendoza, and Carlos A. Vegh. 2013. “How Big (Small?) Are Fiscal
Multipliers? Journal of Monetary Economics, 60 (2013): 239–254.
International Monetary Fund. 2010. World Economic Outlook: Recovery, Risk and Rebalancing,
Chapter 3. Washington, DC.
Leeper, Eric M., Alexander W. Richter, and Todd B. Walker. 2012. “Quantitative Effects of Fiscal
Foresight.” American Economic Journal: Economic Policy, 4 (2): 115–144.
16
Leeper, Eric M., Michael Plante, and Nora Traum. 2010. “Dynamics of Fiscal Financing in the
United States.” Journal of Econometrics, 156 (2): 304–321.
Leeper, Eric M., Nora Traum, and Todd B. Walker. 2011. “Clearing Up the Fiscal Multiplier
Morass.” National Bureau of Economic Research Working Paper, No. 17444.
Leeper, Eric M., Todd B. Walker, and Shu-Chun Susan Yang. 2009. “Government Investment and
Fiscal Stimulus in the Short and Long Runs.” National Bureau of Economic Research Working
Paper, No. 15153.
Michaillat, Pascal. 2012. “Fiscal Multipliers Over the Business Cycle,” CEP Discussion Paper, No.
1115. The London School of Economics and Political Science.
Owyang, Michael T., Valerie A. Ramey, and Sarah Zubairy. 2013. “Are Government Spending
Multipliers Greater During Periods of Slack? Evidence from Twentieth-Century Historical Data,”
American Economic Review: Papers and Proceedings, 103 (3): 129–134.
Parker, Jonathan A. 2011. “On Measuring the Effects of Fiscal Policy in Recessions.” Journal of
Economic Literature, 49 (3): 703–718.
Perotti, Roberto. 2004. “Public Investment: Another (Different) Look.” Innocenzo Gasparini
Institute for Economic Research Working Paper, No. 277.
Perotti, Roberto. 2011. “The ‘Austerity Myth’: Gain Without Pain?” National Bureau of Economic
Research Working Paper, No. 17571.
Ramey, Valerie A. 2011a. “Identifying Government Spending Shocks: It’s All in the Timing.” The
Quarterly Journal of Economics, 126 (1): 1–50.
Ramey, Valerie A. 2011b. “Can Government Purchases Stimulate the Economy?” Journal of
Economic Literature, 49 (3): 673–685.
Ramey, Valerie A. and Matthew D. Shapiro. 1998. “Costly Capital Reallocation and the Effects of
Government Spending.” Carnegie-Rochester Conference Series on Public Policy, 48 (1): 145–194.
Ramey, Valerie A. and Sarah Zubairy. 2014. “Government Spending Multipliers in Good Times
and Bad: Evidence from U.S. Historical Experience.” Working Paper (November 22, 2014),
http://tinyurl.com/lcxkj2c.
Reichling, Felix and Charles Whalen. 2012. Assessing the Short-Term Effects on Output of
Changes in Federal Fiscal Policies. Congressional Budget Office Working Paper, No. 2012-08.
Riera-Crichton, Daniel, Carlos A. Vegh, and Guillermo Vuletin. 2012. “Tax Multipliers: Pitfalls in
Measurement and Identification.” National Bureau of Economic Research Working Paper, No.
18497.
Riera-Crichton, Daniel, Carlos A. Vegh, and Guillermo Vuletin. 2014. “Procyclical and
Countercyclical Fiscal Multipliers: Evidence From OECD Countries.” National Bureau of
Economic Research Working Paper, No. 20533.
17
Romer, Christina D. and Jared Bernstein. 2009. “The Job Impact of the American Recovery and
Reinvestment Plan.”
Romer, Christina D. and David H. Romer. 2010. “The Macroeconomic Effects of Tax Changes:
Estimates Based on a New Measure of Fiscal Shocks.” American Economic Review, 100 (3): 763–
801.
Sutherland, Alan. 1997. “Fiscal Crises and Aggregate Demand: Can High Public Debt Reverse the
Effects of Fiscal Policy?” Journal of Public Economics, 65 (2): 147–162.
Taylor, John B. 2011. “An Empirical Analysis of the Revival of Fiscal Activism in the 2000s.”
Journal of Economic Literature, 49 (3): 686–702.
Van Brusselen, Patrick. 2009. “Financial Stabilisation Plans and the Outlook for the World
Economy.” European Network of Economic Policy Research Institutes Working Paper, No. 55.
Woodford, Michael. 2011. “Simple Analytics of the Government Expenditure Multiplier.”
American Economic Journal: Macroeconomics, 3 (1): 1–35.
Zandi, Mark. 2011. “At Last, the U.S. Begins a Serious Fiscal Debate.” Moody’s Analytics,
http://tinyurl.com/mxh62y6.