A comparison of estimators of output multipliers from incomplete input-output data

Abstract

This paper presents derivations and empirical tests of three formulas for estimating output multipliers with only limited inputoutput type data. The first formula assumes only knowledge of the proportion of the value of output of the firm or industry in question which is respent within the region, as well as an estimate of the average of this value over all industries. The second formula assumes this value as well as the relative values of the I-O "row totals." The third equation assumes data equivalent to a column in the I-O matrix is known for the specific firm or industry in question. The data requirements for all three formulas are modest but are more for (2) than for (I) and are more for (3) than for (2). All three formulas are shown to be quite reliable but the third gives especially close estimates of the actual multipliers for the several models tested. It is concluded that fairly rough estimates of the average interindustry proportion can be safely used but it is important to obtain the firm or industry expenditure (direct effects) data as accurately as possible.

Shortcut ‘Input-Output’ Multipliers, Alive and Well: Response to Jensen and Hewings

“…As a preliminary remark, we should like to point out that Jensen and Hewings (1985) have based most of their comments on our papers that were written between 1977 and 1981 (Burford and Katz, 1977a;1977b;1978;Katz and Burford, 1981a;1982). The tone and and content of their recent note is quite different to that in an earlier paper (Jensen, 1978).…”

Shortcut formulas for output, income and employment multipliers

“…Therefore the mean or expected value of the distribution of each coefficient random variable aij follows from the following derivation (also see [8]). …”

“…The "random matrix approach" is an alternate method of producing shortcut multiplier formulas which allows greater flexibility in the construction of formulas under varying amounts of information, and also provides a theoretical framework to analyze the sensitivity of these shortcut multiplier formulas. This statistical technique was introduced by Katz and Burford [8] to derive three different shortcut output multiplier formulas under varying levels of known information about the economic region. The purpose of this paper is to extend the application of the "random matrix approach" to the generation of not only output but also income and employment multipliers.…”

“…The shortcut multiplier formula technique is another approach that is based on the idea that not every input-output coefficient is important in the formation of a particular industry's multipliers, but rather there are several parameters that critically affect the size of each multiplier (see for example Burford and Katz [4], Katz and Burford [8]). This method attempts to construct a formula using these "critical parameters" which represent only a small part of the data required by the actual input-output matrix.…”

Shortcut Multiplier Formulas for Interregional Input-Output Models

“…Shortcut multiplier formulas provide an alternate approach to the measurement of interregional impacts. Burford and Katz (1981) (see also Katz and Burford (1981, 1981a) have constructed shortcut multiplier formulas, based on limited industry detail, that tend to closely approximmate the reported multipliers for various regional input-output models. The shortcut multiplier methodology is based on the theory that several key identifiable elements in the coefficient matrix are critical to the formation of a specific multiplier.…”