Measures

Interactive use of budgets. This measure was built on Simons’ (1994, 1995b, 2000, 2005)

definition and description of interactive control systems where four items derived from the original description of the construct were developed to measure the interactive use of budgets. Since the focus of the analysis on control behaviors in this study was on the interactive use of budgets to uncover emerging market developments, we directed the questions to the HoSM in the respective firms. The respondents were asked to indicate the extent to which the stated use of the budgeting

94

process corresponded to the way things were handled in their firm during 2010-2012. The four items describing the use of budgets were as follows: 1) the budget follow-up process is considered important by top management and they use it continuously, 2) top management often uses budget information to question decisions and discuss ongoing actions with department managers, 3) the budget process is ongoing and demands regular and frequent attention from managers at all levels, and 4) there is a lot of interaction between top management and department managers in the budget process. A seven-point Likert-scale (1=no emphasis; 7=strong emphasis) was applied to assess the items and the construct formed by the items exhibited a Cronbach’s alpha of 0.85, which is considered highly satisfactory (Nunally, 1978). See the appendix for details on the items used in the questionnaire and the respective factor loadings in the exploratory analysis. The measure that was applied to interactive use of budget was the overall mean of the four items.

Strategic planning. The measurement of strategic planning and decentralized

strategy-making was based on existing scales described in the literature and tested in prior studies. Strategic planning was operationalized by the CFO’s assessment of the organization’s emphasis on formal strategic planning using items developed and tested by Boyd and Reuning-Elliott (1998). The items were assessed on a seven-point Likert scale (1=no emphasis; 7=strong emphasis) comparable to applications in previous studies (e.g., Andersen, 2004; Rudd, Greenley, Beatson and Lings, 2008;

Andersen and Nielsen, 2009). In the exploratory factor analysis the items loaded on the planning construct and showed a high Cronbach’s alpha of 0.75, which is considered satisfactory (Nunally, 1978).The mean of the five items was used as a measure for strategic planning.

Decentralized strategy-making. The first aspect of decentralized strategy-making;

‘participation in decision-making’, was assessed by the HoSM on items asking about the degree to

95

which sales managers reporting to the HoSM were involved in the decision-making on five different activities of potential strategic importance. The second dimension; ‘delegation of decision authority’, was assessed by asking about the degree to which sales managers were authorized to make decisions without prior approval from top management on the same set of activities. These items have been validated in previous studies (e.g., Andersen (2000, 2004)). An exploratory factor analysis found that the items loaded on the two decentralized strategy-making constructs, exhibiting Cronbach’s alpha’s of 0.81 for participation and 0.88 for delegation which is considered highly satisfactory (Nunally, 1978). The mean of the five items was used to measure the two dimensions of decentralized strategy-making.

Upside potential. We followed the extant literature on semi-variance and downside risk (e.g.

Belderbos, Tong, and Wu, 2014; Miller and Leiblein, 1996; Tong and Reuer, 2007) and constructed a corresponding measure of upside potential determined as the second-root upper partial moment (Fishburn, 1977):

ܷ݌ݏ݅݀݁݌݋ݐ݁݊ݐ݈݅ܽ ൌ ට^{σሺோை஺೟ିூோை஺}_௡ ^೟ሻమ ; for ROAt > IROAt

– ൌ ʹͲͳͲǡʹͲͳͳǡʹͲͳʹǡʹͲͳ͵

We followed the inclination of behavioral decision theory where executives exhibit a tendency to see risk as a failure to meet a certain aspirational performance level (Mao, 1970).

Behavioral studies find that firms tend to use average industry performance as reference points (Frecka and Lee, 1983; Lee and Wu, 1988). In line with Miller and Reuer (1996) we assumed that firms adapt their target level (IROA) on an annual basis computing the upside potential based on return on assets (ROA) as the relevant measure of performance. That is, if a firm is underperforming its peers, upside potential is equal to zero. If the firm is outperforming its peers,

96

the firm has a dispersion of its ROA relative to its target, IROA. A four-year period, 2010-2013, was chosen to obtain data for the construct because besides from corresponding to the period addressed by the questionnaire it provides a time lag in the dependent variable. Ascribing the process constructs and performance outcomes to the same time period assumes that the processes and their resulting effects are contemporaneous over the measured time periods and thereby reduces the possibility of reverse causality. The measure of upside potential de facto captures the probability of over-performance of the firm compared to the closest competitors in the industry during the four-year period.

Control variables. Past research has found that industry context can have a significant

influence on general performance and the relationship between planning and firm performance (Dess, Ireland, and Hitt, 1990; Miller and Cardinal, 1994). Hence, we used the NACE industry codes as indication of specific industry contexts and to control for industry-related effects.

Furthermore, the management literature has found that firm size can influence both direct performance and the interacting planning–performance relationships (e.g., Khandwalla, 1972;

Lindsay and Rue, 1980). It is also suggested that firm size can affect the choice of management control systems and thereby their relationship to performance outcomes (e.g., Khandwalla, 1972).

Hence, we included firm size measured as the natural logarithm of the number of employees in the firm as a control variable. Additionally, we controlled for both organizational and strategic change by asking the CFO if the organization has made significant structural changes and changes to its strategy in recent years on a 7-point Likert scale. Further, we included diversification and internationalization as control variables by asking the CFO of the percentage share of the firm’s turnover from primary and foreign markets, respectively. As these strategies may make the firm less vulnerable to abrupt changes in business or local markets, as they provide access to new businesses

97

or regional markets, diversification and internationalization can have an influence on the firm’s upside potential of performance. Additionally,stock-listing and firm age was included as controls.

Stock-listed firms may have an advantageous access to financial resources that can affect the ability to exploit market opportunities. Firm age might lower the potential to seize upside potential due to increasing bureaucratization over time as the firm matures.