Copyright Protected © Dr Cillian Ryan 2001

Copyright Protected © Dr Cillian Ryan

Strictly for Econ203 Students only. Feedback Welcome.

Draft: Chapter for

'Encyclopaedia of Economics' Snowden, B. & H. Vane Editors Elgar 2002

Business Cycles : Real Business Cycle Approach

For many years economists have struggled to understand economic fluctuations in the

hope that this would enable them to eliminate the unpalatable effects of recessions and

booms such as unemployment or house price inflation. Under this view of the macro-

economy there existed a desirable level of full- (or natural-level-) employment output

which could be attained save for the vagaries of demand fluctuations. These fluctuations,

according to Keynes, depended on the ‘animal spirits’ of consumers and investors.

Alternatively, according to monetarist, they depended on expectation errors or rigidities

caused by monetary fluctuations. Thus, it was believed that governments, by undertaking

economic research and judicious use of fiscal or monetary policy, could control the worst

excesses of the business cycle.

One of the difficulties with the models used by policy makers to understand the

effect of their policy actions was that the parameters (such as the marginal propensity to

consume) employed in traditional models might change as government economic policy

changed. If this were the case then the policy predictions based on these models would be

useless. This was the so-called Lucas Critique. Lucas instead suggested that models ought

to be based on primary elements of microeconomic behaviour such as utility and

Copyright Protected © Dr Cillian Ryan 2001

production functions. In attempting to formulate such a model Kydland and Prescott

(1984) founded what we now call Real Business Cycle Theory.

Essentially the model they built led them to the conclusion, surprising at the time,

that most fluctuations were caused not by demand variations due to changing

expectations or ‘animal spirits’ but rather by agents reactions to shocks to the production

function or in other words ‘real’ shocks, hence the title Real Business Cycle Theory.

Their ‘discovery’ came about due to the methods they used to build and test their

model. Since the model was not based on reduced form equations (such as the IS or LM

curves we usually estimate) it could not be tested econometrically and instead they used a

technique known as computer simulation. Computer Simulation is sometimes

derogatorily referred to as doing econometrics with one observation! In practice, the

original research took data from a particular base year and using parameters from

previous econometric studies for the consumption, production and other functions fitted

(or calibrated) the model to the base year data. The inter-temporal utility and production

functions were simple Cobb-Douglas type functions, with production being subject to

shocks. These shocks were derived by first calculating the level of economic growth that

could be attributed to changes in labour and capital in any period. The difference

between the predicted change and the actual change (Solow residuals) was attributed to

random economic shocks. The model, which has been calibrated to the base year, is then

run through the computer with a different shock for each year and the resulting choices of
Copyright Protected © Dr Cillian Ryan 2001

consumption, employment, production and investment are noted and compared with the

actual data for the economy.

Kydland & Prescott found that when the computer simulation of their model was

compared with actual US data that the fluctuations in the model economy looked very

like those in the real data. Thus, they concluded, the new model, where fluctuations in

economic variables only arise as a result of shocks to the production function, provided a

good fit of the ‘Stylised Facts’ (see chapter???). This was seen as a dramatic result at that

time, since, as was noted above, it was believed up to that point that production (or

supply-side) shocks only accounted for a very small percentage of the real fluctuations,

and that either demand or monetary shocks were the more important determinants. This

new theory called the traditional view into question.

The policy implications of this were enormous. Instead of the business cycle and

the associated fluctuations in employment, consumption and investment being caused by

some demand-side market failure which the government needed to eliminate, this

research suggested that the fluctuations we observe are optimal responses on the part of

individual consumers, producers and investors to real economic shocks (such as weather,

natural events, strikes, political upheaval and new inventions etc.). It follows that any

attempt on the part of government to smooth out these variations must cause consumer to

deviate from their freely-chosen, optimal actions and hence make them worse off.
Copyright Protected © Dr Cillian Ryan 2001

To put it another way, the Keynesian ideal of striving to attain some target level of

full-employment equilibrium is misplaced. What is actually happening is that equilibrium

output is constantly moving about over time, in response to ‘real’ shocks. Furthermore,

proponents of the Real Business Cycle prefer to think of these shocks not as some

predetermined cyclical process, but as random shocks along the long-run growth path of

the economy

The result also had another profound implication, and along with other research

questioning the efficacy of short-run interventionist policy, has had a significant influence

on many modern western governments. The theory suggested that the object of fiscal

policy was not short-term counter-cyclical intervention, but rather to design tax and

expenditure policies particularly as regards the provision of public goods and education

which maximize the growth rate of the economy. Furthermore, monetary policy has no

role in this framework as its distortionary effects on prices and expectations can only

interfere with optimal consumer choices. Sometimes this policy prescription is referred to

as supply-side economics as it is concerned with devising policies that maximize the

supply of output in the long run. In summary, Real Business Cycle theory suggests that

governments ought to be concerned with achieving the best long-run trend of GNP rather

than being concerned with the fluctuations around it. Needless to say, this view provoked

considerable controversy on a variety of fronts, both because its interpretation of the

process generating the stylised facts we observe was so radically different from our

previous understanding, and because it was seen to diminish the role of government and

was thus seen as a right –wing agenda setting theory (though defenders would argue that
Copyright Protected © Dr Cillian Ryan 2001

the emphasis on designing optimal fiscal policies rather than on discretionary monetary

policy is strongly in the Keynesian tradition). The theory was thus attacked on a number

of levels.

Economists first tackled the claim that the model fitted the data well. It was

suggested that the Cobb-Douglas forms of the utility and production functions employed

by researchers were too convenient, that their choice of parameters for these functions

based on past empirical work was arbitrary, and unlike econometric studies, the claim

that the results of the simulated model were ‘close’ to the real data was purely subjective

and unverifiable.

Subsequent refinements of the simulation technique (which had been employed in

international trade and tax policy studies for decades) helped to address some of these

problems. An advanced econometric technique, known as the Generalized Method of

Moments, allows the data to tell the researcher what the appropriate functional form and

parameters of the functions should be and removes some of the subjective choices a

researcher must make when setting up the model. Similarly, sensitivity tests can be

conducted to see just how much the results of the model vary as the parameters of the

functions vary. Finally, it was noted that the stylised facts we are measuring the model

against are not rigid observations but rather they are derived by sampling real economic

data, and we can measure the sampling variability. Thus it is possible to ask if a given set

of simulated observations is likely to have come from the underlying population given

the real data at our disposal. Despite these refinements to the simulation method, there are
Copyright Protected © Dr Cillian Ryan 2001

many economists who remain sceptical about this type of approach, advocating instead

the refinement and improvement of econometrics-based macroeconomic modelling.

The second key development was to marry some of the improvements in the

methodology, as outlined above, with refinements to the model to tackle areas where the

simulation results did not perform well in matching the real data. The key deficiency was

the performance of the model with respect to the labour market. Firstly the original

simulated model could only generate about half the fluctuations actually observed in the

employment market, and secondly the model failed the famous ‘Dunlop-Tarshis’ test.

This is the stylised fact, established as long ago as the 1930s, that there was a near zero

correlation between hours worked and real wages over the cycle.

One line of research tried to improve the models ability to match labour market

variations by allowing for, amongst others, such phenomenon as the effect for a minimum

40 hour week on the employment/unemployment /over-time/rest decision and the effect

of house-hold opportunities (doing your own house-work/maintenance/decorating versus

working and paying someone else). These sorts of refinements all succeeded in

substantially improving the performance of the model.

The second line of research tackled the ‘Dunlop-Tarshis’ problem that there

should not be any relationship between wages and hours. The real business cycle model

implies a positive correlation because the demand curve for labour moves up and down

the labour-supply curve as shocks hit the production function. By contrast the older
Copyright Protected © Dr Cillian Ryan 2001

Keynesian/Monetarist approach saw the labour supply curve as moving up and down the

labour demand curve (for example, due to employees mis-perceptions of the real wage or

sticky wage contracts) and thus predicts a negative correlation. Thus what was necessary

was a model that would generate movements in both the labour demand and supply

curves. This was addressed by, for example, allowing for the influence of cyclical

government spending on labour supply decisions and by extending the model to

incorporate Keynesian-type labour market failures.

More recently, research in this field has concentrated in trying to extend the model

to explain nominal and monetary stylised facts rather than the real issues. In these models

either money or a banking sector is incorporated into the model and then the influence of

either real production or nominal money supply side shocks are considered. To date these

models have enjoyed mixed success in explaining the monetary stylised facts and much

work remains in this area.

Bibliography

Christiano, L. J. and M. Eichenbaum (1992), ‘Currency, Real-Business Cycle Theories

and Aggregate Labor Market Fluctuations’, American Economic Review, 82, June

Danthine, J.P. and J.B. Donaldson (1990), Efficiency Wages and Business Cycle Puzzle’,

European Economic Review, 34, November.

Copyright Protected © Dr Cillian Ryan 2001

Kydland, F. E. and E. C. Prescott (1982), ‘Time to build and Aggregate Fluctuations’,

Econometrica, 50, November.

Long, J. B, Jr, and C. I. Plosser (1983) ‘Real Business Cycles’, Journal of Political

Economy, 91, February.

Ryan, C and A.W. Mullineux (1997), ‘The Ups and Downs of Modern Business Cycle

Theory, in Reflections on the Development of Modern Macroeconomics in B.

Snowden B and H.Vane (eds), Aldershot : Edward Elgar,