cut-off point or cutting interval for any scale variable.3As mentioned above, the dedication of too much effort to the quest for a comprehen- sive definition of smallness may not be very useful, especially when tra- ditional concepts are at stake. We, therefore, provide only a short over- view of the most important characteristics for possible definitions and demonstrate that traditional indicators are not very helpful in a theore- tical context, before turning to a more promising approach to defining smallness of countries. Nonetheless, for empirical studies, it often suffices to employ trivi- al definitions of smallness or size. The use of simple one-dimensional measures like population, area or GDP as proxies for size is often ade- quate. It will, however, be shown in Chapter 3 that in some cases the re- sults are not independent of the choice of size 
proxy. 2.1.1 Size and population There is more to the definition of smallness than population, but it is clear that population is the most important proxy for country size. According to Olafsson (1998, p. 8) the number of inhabitants deter mines two eminent characteristics of an economy: – the size of the internal market – the possible degree of specialization within a country Clearly, international trade and the extent of international division of la- bor are two other important factors in determining a country’s «real» or «effective» (economic) size, but there are some important economic cha- racteristics which are mainly determined by the number of inhabitants, like the degree of international representation or the structure of the po- litical system. Table A.1 shows different arbitrarily chosen cutting points for the formation of a set of small countries (based on population); none of them has a theoretical rationale, but all are underpinned by intuitive considerations. Table 2.1 gives an overview of the smallest countries in the world judged by the number of inhabitants. 18Smallness 
of countries: concepts and definitions 3Note that in the empirical parts of this study we sometimes rely on large country samples which include countries of all sizes, in order to assess the effects of size more generally. These parts of the study do not require a cutting point at all.