International Max Planck Research School on the Social and Political Constitution of the Economy
Recounting the Beans: The Statistical Construction of Fiscal Reality
Studies on the Social and Political Constitution of the Economy. IMPRS-SPCE, Cologne 2009.
Accounting and statistical frameworks do not just mirror reality. They actually create it. To measure economic "facts", many conceptual decisions have to be taken.
While those framing decisions may be consensual among experts, a lot of them are not, or rely on arbitrary or ambiguous grounds. A closer, inductive look into the
measurers' realm reveals that economic reality is actively constructed - either by mere convention or by deliberate decisions.
This setting becomes highly important if we take the institutional context into account, in which those numbers are produced and used. Representatives of the involved
measurement professions (i.e. economic statisticians or financial accountants) and national and international organizations strive to effectively deal with the upcoming
measurement questions. It turns out, they try to actively find justifications for critical treatments. Trust in numbers is actively managed with the promise to solve
open questions in a technical manner. This goes hand in hand with many data users' interest to blindly rely on numerical facts. Yet an analysis of the debates that lead
to the issuance of measurement standards clearly shows, that the suggested precision and neutrality is a myth.
This has important implications, as economic numbers are increasingly used to rationalize decisions in the fiscal, economic and business environment. By not being aware
of the constructive elements of economic measurement, decisions are effectively either delegated to certain expert communities or are essentially influenced by accident.
This study looks at the process of the statistical creation of fiscal reality from and end-to-end perspective. In the field of international fiscal statistical standards
(i.e. the IMF Government Finance Statistics framework and the related System of National Accounts) the in practice relevant dimensions of reality construction are first
derived conceptually. Then one debate by standard setters on a specific measurement issue is reconstructed and the spectrum of debated measurement options is derived. In a
third step, a semi-empirical dataset is built, that operationalizes the alternative measurement conventions respectively. This way the numerical impact of the discussed
options can be shown. These alternative data are then eymployed to undertake an econometric sensitivity analysis of a standard fiscal impact projection model.
The results suggest that conceptually, fiscal reality is constructed in many important dimensions. Second, the analysis of the example debate reveals that the decisions
of measurers are at the same time both over- and underdetermined by the justification principles used in practice. Thus significant room of maneuver exists when deciding
on statistical standards. Thirdly, the numerical implications of the largely arbitrary or at least ambiguous solution spectrum are profound. The results derived from the
econometric analysis show dramatically different results, depending on how fiscal reality is measured. Together these results show that we dramatically err if we treat
economic numbers as a technical area. Rather we need to profoundly understand the social context in which the numbers come into being.