creasing pressures for more dynamism and rapid globalization. However, we do feel
confident that these data are suggestive of the complexity of the performance evalu-
ations problem faced by financial managers in global corporations. First, it seems
quite clear that performance measurement systems are being allowed to adapt to local
conditions. At some sites, total liabilities were the focus of quantitative goals for per-
formance; at other sites, sales of particular products, such as life insurance or mort-
gage products, were the focus. At another site, actual behaviors like numbers of
manned calls were the targets; while at yet another, total costs were the item con-
trolled. We have no doubt that there are conditions under which it makes perfect
sense to allow performance monitoring to adapt to what is important in the local en-
vironments of various sites in a global company. However, it is also true that com-
bining these various measures of performance into a coherent vision, as the advocates
of dynamic performance measures would suggest is necessary, is likely to be a fairly
complicated process.
Even from the point of view of traditional management accounting concerns, the
overall performance picture is likely clouded by this multitude of measures. The pa-
rameters by which goals for sales of products, goals for profitability, and goals for
total costs adjust over time are quite distinct. Similarly, the process that seems to gov-
ern the setting of goals for units and personnel at the technical core seems to be quite
distinct from the process that governs the setting of goals for boundary spanners. We
have no doubt that clever budgets can get all these numbers to balance, even after the
myriad of currency translations and other considerations imposed by the purely
global aspects of the transactions. However, it is not at all clear that balancing the
numbers and eliciting the best performance need to or even can coincide. Further-
more, as experiments with feedback in simple supply chains have shown, misper-
ceptions of inputs in dynamic decision-making contexts is pervasive.^44 The stringing
together of the results of operations using a variety of performance measures from a
variety of locations creates complexities that may very well have similar effects. Fur-
ther analysis of the experience of multinational companies as they try to globalize
their increasingly dynamic performance measurement systems is clearly in order.
While the results reported here are far too preliminary to provide conclusive evidence
concerning the scope of the problem, they are certainly suggestive of the need to con-
26 • 12 DYNAMIC PERFORMANCE MEASUREMENT SYSTEMS FOR A GLOBAL WORLDMeasure:Annual branch goals for life insurance products stated in local currency.
Data:102 usable observations were obtained from 2 years of data on 102 branches.
Predictor Variable Coefficient
Constant .938
Log of Previous Aspiration Level .404
Log of Previous Performance Level –.396*
Branch Conversion .103
p < 0.05 **p < 0.01
Restated Model:
LogGoalt= .938 + .800 LogGoalt–1+ .396 (LogPerft–1– LogGoalt–1) + .103 BrConv
Exhibit 26.6. Goals for Other Sales (boundary spanning units).
(^44) Sterman, 1989.