Measuring… What? Notes on Some Globalization Indices

Abstract

The article examines some of the instruments devised to measure globalization, in particular the CSGR Globalisation Index and the A.T. Kearney/Foreign Policy Globalization Index. The article begins by comparing the features and results of these instruments, and then highlights their strengths and weaknesses. The two most significant weaknesses are as follows. The first consists in what Beck has called ‘methodological nationalism’: the fact that attempts are made to study an essentially transnational phenomenon on a national basis. The second weakness resides in the problems—in terms of the completeness, adequacy, and timeliness of information—of the database used to calculated these indices. Also to be emphasized is that analysis of the strengths and weaknesses of these indices cannot be separated from the purpose for which they have been devised. In other words, there is no one ‘best’ instrument in absolute terms; rather there is an array of instruments with which to pursue particular knowledge goals. Finally, discussion of the techniques used to measure the phenomenon provides an occasion to raise considerations concerning the nature of globalization.

A Chinese version of this article's abstract is available online at: www.informaworld.com/rglo

El artículo examina algunos de los instrumentos diseñados para medir la globalización, en particular el Índice de Globalización CSGR (por sus siglas en inglés), y el Índice realizado por A.T. Kearney y la publicación de política exterior (Foreign Policy Magazine). El artículo comienza con una comparación entre las características y los resultados de estos instrumentos, y luego destaca sus ventajas y desventajas. Las dos desventajas más significantes son las siguientes. La primera consiste en lo que Beck ha llamado el ‘nacionalismo metodológico’: el hecho de que haya habido intentos para estudiar un fenómeno esencialmente trasnacional a partir de una base nacional. La segunda desventaja reside en los problemas—en términos de compleción, adecuación y actualidad de la información—de la base de datos usada para calcular estos índices. También hay que resaltar el hecho de que el análisis de las ventajas y desventajas de estos índices no puede separarse del propósito para el cual han sido diseñados. En otras palabras, no existe un instrumento ‘mejor’ en términos absolutos; sino más bien una serie de instrumentos con la que se persigue metas de conocimiento particular. Finalmente, la discusión de las técnicas usadas para medir el fenómeno, brinda una ocasión para plantear las consideraciones concernientes a la clase de globalización.

Notes

A Chinese version of this article's abstract is available online at: www.informaworld.com/rglo

Following a celebrated article by T. Levitt (1983). Not to be overlooked, however, are pioneering studies by authors belonging to other disciplinary fields, most notably Nettl and Robertson (1968).

Besides this mainstream view, it is still widely believed that the economic aspects of globalization constitute its most advanced dimension and therefore drive the process. To be pointed out, however, is the decidedly contrary opinion of M. Waters that it is instead the cultural dimension of globalization which is its engine: ‘material exchanges localize; political exchanges internationalize; and symbolic exchanges globalize. It follows that the globalization of human society is contingent on the extent to which cultural arrangements are effective relative to economic and political arrangements. We can expect the economy and the polity to be globalized to the extent that they are culturalized, that is, to the extent that the exchanges that take place within them are accomplished symbolically. We would also expect that the degree of globalization is greater in the cultural arena than either of the other two.’ (1995, pp. 9–10; emphasis in the original).

Another brief overview of the contrasting interpretations of globalization processes in the sociological literature is provided by Fiss and Hirsch (2005, p. 32). For more thorough surveys of the topic, see Sklair (1999) and Guillén (2001).

By ‘indicator’ is meant a specific, empirically measurable, concept able to furnish information about a more general concept which is not empirically measurable (Corbetta 1999, p. 115).

When selection is made of the indicators for each of the dimensions identified, a balance must be struck between two criteria: (a) optimal representativeness with respect to the dimension considered; (b) availability, quality, timeliness, and cost of the corresponding information.

Without specifications for each of the points that follow, these are the texts referred to here to identify the desirable features of an index constructed to measure a complex social phenomenon: Drewnowski (1970), Graziosi (1979), Morris (1979), Cipolla (1987), United Nations (1989), Church and McHarry (1994), Alberti et al. (1995), Scamuzzi (1996), Scidà (1997), Cartwright (2000), UNDP (2005).

Available at: www.atkearney.com

The version published in 2006 is identical to the one of 2005. This is the first time that the index has not been modified from one year to the next.

The authors of the index acknowledge that these dimensions capture only some aspects of globalization, and that it would be appropriate to include cultural exchanges as well. This is not done, however, because of the lack of reliable data on this dimension (Foreign Policy, 2003: 63).

That is, the maximum value on the basis of which the normalization is performed varies from year to year for each variable. Previously, only one maximum value (and the minimum value, now not considered) was used for normalization and corresponded to the highest (and the lowest) of all those recorded for the variable since 1998.

The problem is that, for each variable, the maximum value from year to year may refer to different countries. Yet information on how this ‘scale factor’ is calculated has not been published. Is a reference country taken as the benchmark, or is recalculation made of all the ‘scale factors’ on the basis of the country which, at that particular moment in time, records the highest value for that particular variable? Also to be pointed out is that, because this procedure is subsequent to normalization on the scale 0–1, it may unduly increase the effective weights in the overall index of the factors for which substantial growth has been recorded in recent years, for example those relative to the technological dimension. Indeed, the United States is given high rankings by the A.T. Kearney/Foreign Policy Globalization Index precisely because of its good performance on the technological dimension (year of reference 2004), although the latter nominally accounts for just 10% of the overall value of the index.

Available at: www.csgr.org

Some criticisms of the WMRC G-Index are made by Martens and Zywietz (2006).

Dreher's position is even more untenable if we consider that he measures this dimension more concretely by using the number of McDonald's restaurants on the national territory as his indicator. But why, one asks, should one not instead measure the level of cultural globalization in terms of the number of Chinese restaurants or Italian pizzerias?

Proposals which merely modify the system whereby weights are attributed to the various variables considered in the A.T. Kearney/Foreign Policy Globalization Index.

Where the figure for such a long time interval is available. If the interval considered for the normalization is not specified, one may presume that it is the maximum interval for which the figure is available, or else the authors may have resorted to an estimate.

Using the well-known formula: normalized value = (observed value – minimum value)/(maximum value – minimum value).

As the authors themselves acknowledge, ‘panel normalisation has both advantages and disadvantages. The advantage is that with panel-normalised data, we can make meaningful comparison over time for a given country or indeed between countries. A disadvantage, discussed in detail in Lockwood (2004), is that when additional years of data are added to the database, the maximum or minimum value of a variable may change, and those variables affected then have to be re-normalised’. This problem can be solved by fixing, on the basis of past observations and predictions for the future, of minimum and maximum invariable thresholds. However, in its turn, this solution has the drawback of identifying a situation of maximum possible globalization, which seems to conflict with the profoundly dynamic nature of a process whose future outcomes at present seem difficult to predict in full.

For technical details on this procedure see Lockwood and Redoano (2005).

Equally arbitrary (and therefore criticizable) is the solution adopted by Martens and Zywietz (2006), whose globalization index assigns equal weight to all the indicators considered.

Note that, following a recent updating of the data, the weights attributed to the various indicators do not seem to have been revised.

For technical details on this regression see Lockwood and Redoano (2005).

Although the authors have been repeatedly questioned on this point, they have not furnished explanations.

When possible, the missing data are estimated by means of a linear interpolation procedure.

Also confirming the presence of this dynamic is the fact that the WMRC G-Index—cited above—ranks Liechtenstein as the most globalized country in the world (Randolph, 2001). Liechtenstein does not appear in the lists of either the CSGR Globalisation Index or the A.T. Kearney/Foreign Policy Globalization Index.

However, the proof would be much more convincing if these were two instruments with radically different structures.

The definitions of GDP and GNP are not reported here, assuming that they are sufficiently well known. To be noted only is that these two indicators are largely interchangeable in the literature. Scidà (1997) has pointed out that whereas GNP was initially preferred, GDP is now more widely used.

See, for example, Drewnowski (1972), Seers (1972), Morris (1979), Horn (1993), Parfit (1993), Streeten (1995), Sen (1999), Gallino (2000), Scidà (2004).

For a detailed examination of the history of development measures and a description of all the instruments mentioned in this section, see Caselli (2001).