Teaching Modern Macroecsnornics at the Principles Level

Ideas taught at the Macroeconomics Princi- learning, experiments, or the use of new media.'
ples level should satisfy two goals. First, they The teaching ideas are similar to those used by
should be simple enough to be both understand- David Romer (2000) and Taylor (1998). but the
able and memorable for the beginning student. purpose of this payer is to show how closely
Second, they should be consistent both with the they represent modeni macroeconomics.
modern economy and with the macroeconomic
models of this econonly that are used in practice I. What Is Modern Macsaecarnomlcs?
for policy evaluation. There is no necessary
conflict between these two goals. The greater At the broadest level, I think it is useful ro
the consistency between the ideas taught in the emphasize five key components of moderf:
classroom and the models used in practice. the macroecor~o~nics (see Taylor, 1997). First, ihe
easier the ideas are to understand and the wor- long-run real GDP trend, or potential CDP, can
thier they are of being remembered. be understood using the growth model that was
It would be an exaggeration to say that a first developed by Robert Solow and that has
consensus now exists in advanced research now been extended to make "technology"
about how macroeconomics should evolve in explicitly endogenous, Second. there is no Iong
the future. Debates continue, for example, run trade-off between inflation and unemploy-
about the usefulness of models with represen- ment, so that monetary policy afkcls inflation
tative agents or what it means to have a fully but is otherwise neutral with respect to real
articulated model of money. Nevertheless, at variables in the long run. Third, there is a short-
the practical level, a comnlon view of macro- run trade-off between inflation and unemaslov- 2 d

economics is now pervasive in policy- ment with significant imnlplications for economic
research projects at universities and central fluctuations around the trend of potential GEP;
banks around the world. This view evolved the trade-off is due largely to temporarily sticky
gradually since the rational-expectations rev- prices and wages. Fourth, expectations of ir18a-
olution of the 1970's and has solidified during tion and of future policy decisions are endoge-
the 1990's. It differs from past views, and it nous and quantitatively significant. Fifth,
explains the growtln and fluctuations of the m ~ n e t a r ~ - ~decisions
o l ~ c ~ a-re best thought of
modern economy; it can thus be said to rep- as n-ules, or reaction functions. in vl!hich tkie
resent a modern view of macroeconon~ics.' short-term nominal interest rate (the instmrraent
The purpose of this paper is to show how this of policy) is adjusted in reaction to cconornis
modern macroeconomics can be taught at the events.
Principles level. I focus on macroeconomic con-
cepts (including economic growth and fluctua-
tions) and their graphical representation rather
than on techniques of delivery, whether active The first and second points suggest t h b
teaching beginning rtudents the Soloe*?model,
augmented with endogenous technology, is the
* Department of Economics, Stanford University, Stan- first step toward teaching them modern macro
ford, CA 94305. I am grateful to Michael Salemi and econo~nics.But how much of that model is
Marcelo Clerici-Arias for helpful comments.
'The preface and many of the papers in Taylor and i~~anageab'le by students at the Principles level"
Michael Woodford (1999) describe such an evolution. Var- Bn my \lienr, the simple growth accounting for-
ious names have been suggested for the n~acroeconomic
approach that is now so common including "new neoclas-
sical synthesis" (Marvin Goodfriend and Robert King,
1997) or "new Keynesian economics" (Richard Clarida et 'Ideas {or lectures in introductory economics are dis-
al., 1999). cussed in Marcelo Clerici-Arias and 'Taylor (2000).
VOL. 90 NO. 2 TEACHING MACROECONOiMIC PRINCIPLES 91

mula relating labor productivity growth to the monetary-policy evaluation used this type of
growth in capital per worker and to the growth model (see Taylor, 1999). Models now used for
in technology should be the center of the dis- policy evaluation at the Federal Reserve, the
cussion. Tlying to explain the steady-state European Central Bank, the Bank of Canada,
growth equilibrium is too abstract for beginning the Bank of England, the Reserve Bank of New
students and is better left for more advanced Zealand, and the Central Bank of Brazil also fall
courses. In my experience it is straightforward into this category.
for students to use the growth accounting for- Sorne of these models (such as that of Julio
mula. They enjoy using it to explain why Rotemberg and Woodford [ I 9971 or Lars E. 0.
growth slowed down in the 1970's or to deter- Svensson [2000]) are more explicitly tied in
mine whether the pickup of growth since the with microeconomic foundations than others.
mid 1990's in the United States (a key feature of Some of the models are very small, with only
the "new economy") is due to more capital or to three equations (such as that of Jeffrey C. Fuh-
better technology. Students sometimes find that rer and Brian F. Madigan [1997]), and some
the formula is too mechanical, and providing an have many equations. But all the models can be
intuitive explanation of it is helpful. (One could boiled down to three relationships and three
derive the formula graphically using shifts and variables:
movements along a production function, or (i) 'The first relationship is between real GDP
present a Cobb-Douglas production function and the real interest rate. Such a relationship can
and compute total factor productivity using data be explained intuitively to Principles students,
on labor and capital, but these raise the level of but I can use some algebra here. The simplest
difficulty for many students and are probably algebraic form would be y = -ar + u , where
better left for more advanced courses.) The y is real CDP (measured relative to the potential
technology term in the growth accounting GDP that comes from the Solow growth
formula is useful for focusing attention (2 la model), r is the real interest rate, and u is a shift
endogenous-growth theory) onthe determinants term such as a shock to exports or fiscal policy.
of technology growth, including education, re- This relationship is analogous to an IS curve. It
search and development, and the process of describes how a higher real interest rate de-
invention and innovation. presses the demand for goods and services in
the economy, The equation can be derived as
the first-order condition of an intertemporal
B. Economic Flt~ctttations maximization problem (see Clarida et al.,
1999). Such a derivation would include a lead
While the growth accounting formula is use- of output on the right-hand side. It can also be
ful for explaining long-term growth in the econ- derived using a Keynesian cross diagram in
omy, other factors (e.g., the short-run trade-off, which the aggregate-expenditure line shifts
expectations, and monetary policy) must be down with a higher interest rate. I find that most
brought into play in order to explain fluctua- students are satisfied with an intuitive explana-
tions of real GDP around the growth trend. tion that higher interest rates discourage invest-
Fortunately there is a simple approach, compa- ment, net exports (because a higher interest rate
rable with the supply and demand model, that raises the exchange rate), and consumption,
can be used to explain these fluctuations in thereby driving down demand.
much the same way they are explained in mod- (ii) The second relationship is between infla-
ern macroeconomic policy research. tion and the real interest rate. The simplest
If one looks carefully at macroeconomic pol- +
algebraic form would be r := brr v , where rr
icy research in the 1990's, one finds a nearly is the inflation rate and v is a shift term. This
universal model being used to explain fluctua- equation is a close approximation to the actual
tions around the growth trend. Many examples behavior of many central banks. When the in-
are found in the papers reviewed in the useful flation rate rises, the central bank takes actions
survey by Clarida et al. (1999). Virtually all of to raise the short-term interest rate (the federal-
the participants in a recent National Bureau of funds rate in the United States) by enough to
Economic Research (NBER) conference on raise the real interest rate (b should be positive);
92 AER PAPERS AIVD PI\'OC'EEDINGS IMAY 2000

this action is aimed at keeping inflation from

rising further and bringing it back down. Cen-
tral banks must decide how milch to raise inter-
est rates in response to inflation, taking the
likely impact on unemployment or real GDP
into account as well. In policy research, other
terms such as real GDP 2re gerierally included
in the policy rule, but a&the Principles level it is
much easier to keep the reaction to one variable.
Observe that this characterization of monetary
policy in terns of the interest rate is different
from earlier Principles treatments where money
is assumed to be fixed or targeted by central
banks; in reality modern central banks make
decisions about the short--terminterest rate, and
much policy research suggests that this is to be iVoies: Starting from an initial equilibrium, a fiscal shock
preferred to a quantity-oriented policy, at least causes the AD curve to shift to the light to AD,, increasing
with current and expected future behavior of real GDP. Thcn the IA curve begins to shift up gradually,
money demand. until thcrc is a new equilibrium. A shift in monetary policy
would be required to bring inflation hack down.
(iii) The t11il.d relationship is between inflation
and real GDB. This is a standxd expectations--
augmented Phillips curve in which the change in
inflation increases when real GDP rises al~ove bIe to construct a diagram that captures the
potential GDP, signaling denland pressures. 'rl:ie essence of the models in a very simple way,
simplest algebraic form of this relationship is n == Combining (i) the real GDPIinterest-rate rela-.
,T- i- cy-I -1 M I where +e is a shift t e n ; thus, tionship and (ii) the central-bank policy rule
inflation rises with a lag when y is greater than gives a ~legativelysloped relationship between
zero. The modern derivation of this equation is inflation and real GDP. This relationship is an
in terms of staggered price-setting by firns with aggregate demand relationship, labeled AD in
some degree of market power. Here again one .Figure I . This relationship can be explained
would expect to find leads of inflation in the intuitively to beginning students, but one can
relationship, so that expectations of inflation derive it algebraically by substituting the alge-
would raise actual inflation (see Clarida et al.,
1999). obtain y -
braic fonn of (i) into the algebraic form of (ii) to
--abri- 4- u av. Movements
along this relationship occur when inflation
(shown on the vertical axis) changes and the
11, A Simple Graphical Wepresentatiolr
central bank changes the real interest rate, caus-
of Economic Fluctuations
ing real GDP (shown on the horizontal axis) to
change. For example, when inflation rises, the
Kn anore advanced courses or in research central bank raises the real interest rate, and this
work, it is of course possible to solve the three causes real GDP to fall. The AD curve shifts to
equations in three unknowns (y, r , and as
rrr) the right if there is a. positive export shock or a
functions of the shocks ( u , v , and w).Thus one fiscal stimulus. Observe that this AD curve is
can investigate how the econonly responds to the relationship between the irlflation rate and
shocks and study normative policy questions real GDP, rather than between the price level
such as how large the coefficient h in the policy and real GDP. 1 have previously labeled this
rule should be. One can even do this wit11 curve ADT, with the ""%'torinflation, to remind
forwasd-looking variables and with rational ex-. teachers that it differs from other aggregate
pectations and consider the three shocks simul-. demand curves; Romer (2000) labels the curve
taneously as random variables. AD, as I have done in Figure I.
At the principles ievel, however, we need a Relationship (iii) can also be represented in
much sin~plerapproach. Fortunately it is possi,-. Figure 1 . It is a flat line, labeled IA. The IA line
VOL. 90 NO. 2 TEACHING MACROECONOMIC PRINCIPLES 93

shifts up over time when real GDP is above flicts or resource constraints), then it is impor-
potential GDP and shifts down over time when tant for the macro course to spend some serious
real GDP is below potential GDP. The line time covering key micro principles.
would be upward-sloping if current real GDP An alternative is to offer a one-term introduc-
rather than lagged real GDP affected inflation, tory course with microeconomics coming be-
but the flat case is realistic and easier for stu- fore macroeconomics. This is the way elementary
dents. Because this line represents the slow ad- economics is taught at Stanford, and I think the
justment of prices or of the inflation rate, it simple approach to teaching modem macroeco-
could be called either the price-adjustment (PA) nomics outlined here helps make a one-term
line or inflation-adjustment (IA) line. The line course work. But many faculty members feel
takes the place of the aggregate-supply relation- that there is too much economics to teach in one
ship in AD-AS treatments. Following Romer term.
(2000), I label it IA in Figure 1.
Real output and inflation are determined at IV. Conclutrion
the intersection of the AD and IA curves in
Figure 1. How does one explain the effect of a In this paper I have argued that there is a
demand shock or an inflation shock? Suppose distinctive modem form of macroeconomics
there is a fiscal stimulus, and suppose it is that is now being used widely in practice, even
permanent rather than temporary. This stimulus though research on potentially better models
shifts the AD curve to the right, and there is a contimes, and disagreement about the best way
new intersection. GDP rises, but in the short to proceed persists. This theory fits the data well
run, the inflation rate does not rise. Over time, and explains policy decisions and impacts in a
however, the inflation rate does rise, and the IA realistic way. Whether one calls it the "new
line shifts up. The IA line continues to shift up neoclassical synthesis," reminiscent of Paul
until real GDP is back to potential and the Samuelson's original textbook treatment of his
inflation rate is higher. If the central bank "original neoclassical synthesis," or something
wanted to offset this higher inflation rate, then it else, I think it is both appr'opriate and possible
would have to shift the AD curve back down to teach this modern form off macroeconomics at
again. This would cause a slowdown or a reces- the Principles level.3 I recognize that there are
sion as real GDP fell below potential GDP. The many alternative ways to teach macroeconom-
analysis is no more complicated than shifting ics and that what works well for one teacher and
supply and demand curves in elementary micro- his or her students may not be attractive to
economics. And because the inflation rate rather others. I can say that the ideas that I have
than price level is on the vertical axis, there is suggested here have worked well for my
no need to keep shifting the curves up and up students and for me, as well as for others who
and up until they are off the page to describe a have used them.
steady inflation.

It is important to point out that this framework for

111. Micro before Macro? One-Term teaching economic growth and economic fluctuations is
or Two-Term Courses? perfectly consistent with the pre-college Content Standards
endorsed by the American Economic Association's Com-
mittee on Economic Education. The Content Standards in-
What implication does teaching modern mac- clude the ideas that money facilitates econo~nicexchange,
roeconomics in this way have for how Princi- that the interest rate affects investment and saving, that real
income growth is determined by productivity growth, that
ples courses are organized? In my view it investment raises capital and thcbreby raises productivity,
suggests that microeconomics be taught before that unemployment and inflation are costly, and that fiscal
macroeconomics. Certain concepts like treating and monetary policy have impacts on output and inflation
capital and labor as factors of production, or (these are short paraphrases of standards 11, 12, 13, 15, 19,
shifting demand curves around, are probably and 20, respectively). However, the above framework pro-
vides greater specificity and detail appropriate at the college
better understood after some microeconomics. level. In my view, this specificity helps students tie the
If one cannot practically require that micro be content standards together, learn how they are used in
taken before macro (because of scheduling con- practice, and remember them.
94 AEA PAPERS AND PROCEEDINGS MAY 2000

REFERENCES Romer, David. ""Keynesian Macroeconomicv

without the 1,M Curve." Journal of Econornzc
Clarida, Richard; Gali, Jordi and Gertler, Mark. Perspectives, 2000 (forthcoming).
"The Science of Monetary Policy: A New Rotemberg, Julio and Woodford, Michael. "An
Keynesian Perspective." Journal of Eco- Optimization Based Econometric Framework
nomic Literature, December 1999, 37(4), pp. for the Evaluation of Monetary Policy," in
1661-1707. Ben Bernanke and Julio Rotemberg, eds.,
CBerici-Arias, Marcelo and Taylor, John B. "Sur- NBER macroeconomics annual. Cambridge,
prise Side Economics: Ideas for Introductory M A : MIT Press, 1997, pp. 297-346.
Economics Lectures." Unpublished presenta- Svensson, Lars E. 0. '"pen-Economy Inflation
tion at the American Economic Association Targeting." Journal of International Eco
meeting, Boston, MA, January 2000. [Slides nomics, February 2000, $0(1), pp. 155-83.
online: (www.econ1.com).l Taylor, John B. ''A Core of Practical Macroeco-
Fuhrer, Jeffrey C. and Madigan, Brian F. "'Mon. nomics." American Economic Review, May
etary Policy When Interest Rates Are 1997 (Papers and Proceedings), 87(2), pp.
Bounded at Zero." Review o f Economics and 233-35.
Statistics, November 1997,"79(4), pp. 573- _---, economic.^, 2nd Ed. Boston, MA:
85. Houghton-Mifflin, 1998.
Goodfriend, Marvin and King, Robert. "The New ___-, ed. Monetag) policy rules. Chicago:
Neoclassical Synthesis and the Role of Mone- University of Chieago Press, 1999.
tary Policy," in Ben Bernanke and Julio Rotem- Taylor, John B. and Woodford, Michael, eds.
berg, eds., NBER macroeconomics annual. I-landbook of macroeconomics. Amsterdam:
Cambridge, MA: MIT Press, 1997, pp. 23 1- 82. North-Holland, 1999.