Policy Evaluation with Computable General Equilibrium Models

Amedeo Fossati, Wolfgang Wiegard

• Publisher: Routledge
• Year: 2003
• ISBN:
  9781134518807
• Edition: 1
• Topic: Business
• Subtopic: Business General
• Index: Business

Computable General Equilibrium (CGE) modelling is a relatively new field in economics, however, it is rapidly becoming one of the most useful tools for policy evaluation. This book applies CGE modelling to some of the most urgent international economic policy problems, including the Kyoto Protocol, pension reform, and income taxation, and also analyses the methodological issues that arise.

Part I

Methodology

1 Studying the past and future with infinite-horizon simulation models^*

Charles Ballard

1. Introduction

By their very nature, economic models abstract from important aspects of reality. (If models did not employ a degree of abstraction, they would not be models.) And yet, if economists hope to have credibility with each other and with the non-economist majority of the population, it is important that our models not be too far-fetched. The problem of credibility is especially acute for applied general-equilibrium modellers. This is because, by definition, we are dealing with models that are sufficiently complex that they are not amenable to analytical solutions. Thus, we produce results that cannot easily be replicated by other researchers. To the rest of the economics profession and to the rest of the world, it may appear that our results come from a mysterious ‘black box’. Some may take the output of the black box on faith, but others may not. Faith, once lost, is not easily restored. Over the last generation, economists have used a great variety of general-equilibrium simulation models to analyse the effects of tax-policy changes. These include static models, models with overlapping generations of life-cycle consumers, and models in which the consumers have an infinite horizon.

Regardless of the type of model that is used, the goal of virtually all of this research has been to learn about what might occur in the future, if a particular policy change were implemented. This provides the researcher with the comfortable knowledge that he or she is unlikely ever to be proven wrong. In most cases, the authorities never actually implement the exact policy that is considered by the researcher.¹ Even if a given policy is implemented, the tendency of many researchers is to move on to the next paper, without ever asking whether the simulation results conformed to the facts.² It is true that most researchers calibrate their models so that they replicate the original data set, before any policy changes are considered. This does impose a certain amount of discipline on the research. However, in most cases, no attempt is made to validate the results of the model, to see whether it will generate reasonable responses to changes in policy.³

The purpose of this chapter is to explore a different approach to validation and testing of the model. This different approach involves looking into the past, and using the model to analyse the effects of an actual policy change. This new approach could be applied to any kind of model. However, in this chapter, I will focus on the infinite-horizon model. Infinite-horizon models have retained a great deal of popularity, despite their inability to assess issues relating to the inter-generational distribution of wealth. For example, seejudd (1985), Jorgenson and Yun (1990), Lucas (1990), Greenwood and Huffman (1991), Jorgenson and Wilcoxen (1998), Mendoza and Tesar (1998) and many others.

I use an infinite-horizon model to study the effects of the tax-policy changes that were adopted in the United States in 1964–6. To the greatest extent possible, I try to build a model in the way in which one might have been built in 1963, looking forward to the tax-policy changes that would occur in the next few years. I use data for 1962, which became available in the summer of 1963.

The results are somewhat mixed. For some combinations of parameters, the model generates results that are not out of line. However, on the basis of the results presented here, it is not yet possible to draw a truly definitive conclusion regarding the efficacy of infinite-horizon simulation models. This chapter only represents the beginning of an extended research agenda. In future work, it will be necessary to delve more deeply into the problem, by considering, (1) other policy scenarios, (2) other calibration techniques, and (3) different parameter combinations.⁴ However, even at this early stage of the analysis, I believe the results suggest that the economics profession should use caution when interpreting the results of infinite-horizon simulation models.⁵ Researchers who use infinite-horizon models should take great care in specifying the details of the model structure and parameters. My own preference is for the profession to reduce its use of infinite-horizon models, in favour of overlapping-generations models.⁶

2. Tax policy in the United States in the early and mid-1960s

In 1960, John F. Kennedy ran for President on a platform that emphasized the need to revitalize the American economy and society, in order to meet the challenge of global communism. As part of this programme, he eventually submitted to the Congress a package of tax-policy proposals. These proposals led to the passage of a series of changes in tax rates, which were phased in beginning in 1964.⁷

From 1954 to 1963, the marginal tax rates in the Federal individual income tax ranged from 20 to 91 per cent. The rates were then lowered in 1964 and again in 1965. By 1965, the marginal rates ranged from 14 to 70 per cent. Barro and Sahasakul (1983, 1986) estimate that this led to a reduction in the average marginal rate, from 24.7 per cent in 1963 to 21.2 per cent in 1965.

At the same time, the marginal rates in the Federal corporation income tax were also reduced. In 1963, the corporation income tax was levied at a rate of 30 per cent on the first $25,000 of taxable corporate income, and at a rate of 52 per cent on taxable corporate income above $25,000. By 1965, these rates had been reduced to 22 per cent on the first $25,000, and 48 per cent on higher incomes.

While the individual and corporate income-tax rates were being reduced, the Social Security payroll tax was being raised. The combined rates on employers and employees were increased from 6.25 per cent in 1962 to 8.4 per cent in 1966, and the maximum taxable earnings level was raised from $4,800 in 1965 to $6,600 in 1966.

Thus, very roughly, the tax-policy changes of the early and mid-1960s can be viewed as a revenue-neutral set of changes, in which reductions in individual income taxes and corporate income taxes were offset by increases in payroll taxes. In fact, in this chapter, this is the way in which the policy changes will be modelled.

Of course, the tax system is sufficiently complicated that it is never possible to capture every single policy change in a model. For one thing, this chapter will only consider changes in Federal tax policies. We abstract from changes in tax policy at the State and local levels, as well as from changes in the policies of other countries.

Even at the Federal level, we focus only on an important subset of tax-policy changes. For example, after 1963, Federal excise taxes were eliminated on a number of commodities, including excises on automobiles and automobile accessories, cameras, electric appliances, furs, luggage, marijuana, musical instruments, opium, playing cards, phonograph records, photographic film, pool tables, radios, refrigerators and sporting goods other than fishing equipment. We do not consider these tax-policy changes. However, it seems unlikely that these excise-tax reductions would have any substantial effect on the results, because most of these excises raised insignificant amounts of tax revenue. The excise taxes on alcohol, tobacco and gasoline are the o...

Table of contents

• Front Cover
• Policy Evaluation with Computable General Equilibrium Models
• Routledge Applied Economics
• Title Page
• Copyright
• Contents
• List of figures and charts
• List of tables
• List of contributors
• Acknowledgements
• Introduction
• PART I Methodology
• PART II Imperfect competition
• PART III Environment
• PART IV Pension reform
• PART V Miscellaneous
• Index