Accounting-induced performance anxiety: consequences and cures

The anxiety to perform favourably against accounting performance benchmarks has both intended and unintended consequences

by Anne Lillis *

Listen to lecture - September 2008 (MP3 13.5MB)

Introduction

In 1975, Stephen Kerr published a simple but important paper entitled ‘On the folly of rewarding A while hoping for B’, which captured the intended and unintended consequences of performance measures. Kerr identified the ways in which commonly used performance metrics induce behaviour that may enhance reported performance on the measures used, but do not have any substantive performance impact – for example, achieving ‘reject rate’ reduction by stopping quality inspections rather than improving attention to quality. The prominence of accounting metrics in performance evaluation of organisations and their management brings on what I call accounting-induced performance anxiety. The anxiety to perform favourably against accounting performance benchmarks has both intended and unintended consequences.

This lecture examines the phenomenon of accounting-induced performance anxiety by reflecting on three collaborative research projects:

The way intense performance anxiety – for example making losses – affects decision making at the firm level;
The way conventional accounting performance measurement practices within firms can inhibit effective strategy implementation; and
The potential for recent performance-measurement innovations, such as the Balanced Scorecard, to reduce the problem of accounting-induced performance anxiety.

Project 1 – Intense performance anxiety: reporting accounting losses¹

This project examines accounting-induced performance anxiety that derives from capital market pressure to report profits. The accounting profit/loss threshold represents a powerful decision heuristic (empirically derived) because:

The reporting of a loss acts as a trigger for outside intervention (e.g. by boards, regulatory agencies, equity markets and lenders); and
Investors asymmetrically ‘devalue’ a firm in a small loss situation relative to a small profit situation, while these firms are in fact very similar economically.

Although the likelihood of crossing the profit/loss – zero profit – threshold creates severe accounting-induced performance anxiety, this threshold is not a particularly important benchmark in an economic sense. Small profit firms are not ‘profitable’ in an economic sense, as their earnings would not be generating sufficient return on capital.

So if the anxiety was really performance driven, it should be evident somewhere in the small profit area. However, it is the loss threshold which is a relatively arbitrary accounting threshold with little economic meaning.

Managers have a variety of mechanisms available to them to avoid reporting losses. One is ‘earnings management’ – the ability to use the discretion that is available within the reporting of accounting numbers to show the firm in a profit position. It is well documented in the literature that a disproportionately large number of firms report small profits, and a disproportionately small number of firms report small losses. This suggests firms use available mechanisms to avoid the reporting of losses, including the use of accounting discretion to stay on the right side of the threshold. The asymmetric distribution of firms around the zero profit/loss threshold is almost certainly evidence of accounting-induced performance anxiety.

A different reflection of the same type of anxiety arises from the behaviour of firms that have crossed the profit/loss threshold – into a loss position – and want to get back to reporting a profit as quickly as possible. Presumably, these firms have unsuccessfully exploited all available potential for accounting discretion in the bid to report a small profit. More drastic action can include discarding less productive investments that may have been ‘carried’ during more profitable times, or cost cutting in discretionary areas like R&D. Investments in fixed assets are lumpy and irreversible. Reducing the level of employment, on the other hand, provides a continuum (non-lumpy) of divestment potential. Cutting employees turns out to be a highly valuable ‘response’ lever in this situation.

An examination of changes in employee numbers for a very broad cross-section of US firms, classified by their level of earnings deflated by total assets, shows the following patterns:

The average percentage growth in number of employees is systematically lower for loss-making firms than for profit-making firms. There is a significant discontinuity at the threshold and the difference is not explained by differences in economic fundamentals between small loss and small profit firms (Figure 1).²
Firms that cross the threshold from a profit to a loss reduce their investment in employees disproportionately in the year that they cross the threshold.
The effect is more significant in the year following the ‘threshold crossing’ – consistent with firms taking some time to decrease their investment in fixed labour.
In comparing firms that crossed the threshold with firms with a similar earnings-drop that did not cross the threshold, across all quartiles of earnings decreases, the fall in employment is greater for the firms that switched from reporting a profit to reporting a small loss.

Figure 1: Employment changes
Average annual percentage change in no. of employees by firms as a function of different levels of net profit (either side of zero threshold) Figure 1: Employment changes
Source: Pinnuck, M. & Lillis, A.M. 2007, op.cit.

We conclude from this analysis that the profit/loss threshold acts as an anxiety-inducing decision heuristic which has significant economic consequences. The fact that employees are so easily divested – relative to other assets – may render them particularly susceptible to ‘management’ to achieve accounting performance benchmarks.

Accounting-induced performance anxiety within firms

In order to describe the way accounting-induced performance anxiety plays out within firms, consider the following dilemma. A manager of a large division of a local manufacturing company considers recommending the acquisition of a small niche manufacturing company with technology and intellectual capital assets that offer strong potential to enhance the division’s future earnings. This manager is faced with two sets of financial and non-financial information that may convey quite different messages:

Decision-facilitating information – information provided to the manager about the financial and non-financial consequences of the decision. For example: an analysis of the strengths and capabilities of the target company; and discounted cash flow analysis of potential financial consequences of acquisition over a five-year horizon, compared with the ‘do nothing’ option.
Decision-influencing information – information collected by higher level management to evaluate the performance (decision outcomes) of the subunit-managers. For example: divisional profit and return on assets (ROA), measured annually – the basis for managerial performance measurement and bonuses.

The messages may conflict. The discounted cash flow analysis of the acquisition versus the ‘do nothing’ option may support the acquisition, but that decision might ‘depress’ both divisional profit and ROA for the next year or two as the costs of the acquisition are taken up, the asset base is expanded immediately, and the expected payoff is deferred. Which message is ‘right’? Routine annual performance measures will never completely reflect the quality of managerial decisions when measured over short time intervals. Many managerial decisions generate short-term costs and delayed benefits. It is the multi-period analysis that gives the more comprehensive picture of events. Yet we know accounting is always locked into arbitrary reporting cycles.

Which signal will exert the strongest influence on decision-making? Generally, the literature tells us that the decision-influencing information – performance measurement, evaluation, incentives – will dominate. It is difficult to ignore rewarded behaviour. Annual accounting performance metrics within firms tend to induce myopic decision-making and subunit-level optimisation.

Project 2: Customer-responsive manufacturing³

These two related projects examine how performance measurement practices can subvert good strategic decision-making when firms pursue strategies focused on customer responsiveness.

Conventional measures of manufacturing performance focus on efficiency and productivity (e.g. cost variances, scrap, downtime). Conventional structures within manufacturing firms also tend to support mass production at low cost. Efficiency was historically encouraged in manufacturing subunits through task-segregation, by ‘buffering’ manufacturing subunits from the vagaries of markets and customers. Sales subunits were charged with responsibility for dealing with customers. Interdependencies between manufacturing and sales were always high but they were sequential in nature, with the sales and manufacturing interface managed through scheduling. These measures and structures are well suited to manufacturing firms with high levels of product standardisation, stable production processes and a focus on cost minimisation. However, these attributes no longer reflect current strategic priorities in manufacturing.

Figure 2 classifies 36 local manufacturing firms by strategy, and demonstrates the prevalence of multiple strategies focused on quality, service, customer responsiveness and dependability rather than low cost.

Figure 2: Classified summary of strategic orientations. Figure 2 Classified summary of strategic orientations.

The shift from low cost to responsive manufact uring is captured in a quote from a local manu facturer producing heavy-duty men’s work clothes. They “used to do long runs of ‘97 regulars’ or efficient combinations of ‘82 regulars’ and ‘112 stouts’, put them in inventory and sell them.” Now, however, they face demand for a varied mix of products at short lead times. While the product range in this example has remained relatively stable, the firm has significantly increased its responsiveness to customers by shortening lead times and meeting greater within-order variety. These market-initiated changes have increased the rate of production changeovers, reduced batch sizes and increased disruption.

Accounting-induced performance anxiety arises here from the failure of performance measurement approaches to keep pace with shifting strategic priorities. Figure 3 reflects the manufacturing subunit performance measures used by the same 36 firms reflected in Figure 2.

Figure 3: Manufacturing performance measurement practices in the 36 firms reflected in Figure 2. Figure 3 Manufacturing performance measurement practices in the 36 firms reflected in Figure 2.

The paradox here is that the same firms that have moved away from a low cost strategy are using efficiency and productivity measures extensively. Is this the folly of rewarding efficiency while hoping for responsiveness? It raises a challenge for management: how to elicit responsiveness from manufacturing subunits. How do you shift the mindset of efficient lot sizes and maximum throughput to allow for the costs of disruption associated with frequent changeovers, reduced batch sizes and greater product variety? In order to be flexible enough to meet variable customer demands and associated short lead times, manufacturing and sales need to work much more collaboratively than has historically been the case. Interdependencies are described as ‘reciprocal’ rather than sequential, as manufacturing and sales managers negotiate ‘joint’ optimal solutions, rather than manufacturing determining optimal scheduling and efficient batch sizes.

We found that firms deal with this in a couple of ways:

A structural response; or
Reducing the intensity and accounting focus of performance measurement.

Structural response

The structural response involves investing in integrative structural arrangements that facilitate cross-functional co-ordination and interaction – more organic, less mechanistic structures. These structural mechanisms include cross-functional teams, task forces and daily cross-functional meetings. The aim of these devices is to link the efficiency and productivity-focused mindset of manufacturing with the customer-focused mindset in sales, and to facilitate the joint development of optimal production solutions.

Reducing the intensity of the accounting focus of performance measurement

In settings with a strong commitment to customer responsiveness, the absence of standardisation makes it increasingly difficult to specify unambiguous performance standards. Performance standards in manufacturing subunits are generally constructed in the form of standard product costs that specify standard expectations in relation to material, labour and overhead input per unit of output. Such standards are set to reflect a specific efficiency or productivity level (assumed labour standard to assemble a certain number of units per hour). It is notoriously difficult to build the costs of disruption and unpredictable frequent change overs onto standard costs. Firms struggle with this – they are unable to rewrite efficiency standards to incorporate the costs of disruption, but they want manufacturing subunits to be prepared to ‘wear’ the disruption in the interests of customer responsiveness. The way they deal with it is to ‘play down’ the pressure around cost budgets and efficiency standards. In effect, they rely less on accounting performance benchmarks as they would be ‘counter-strategic’.

So what happens when firms do not get this right – when structure and performance measurement impede flexibility?

“We’re nowhere near as flexible as we would like. We always liken it to a battleship where it takes miles to turn it around.”

“Production is so intent on meeting their weekly targets, if a special order comes in they tend to say, ‘Oh no, what a nuisance’, rather than looking at the opportunity presented. And that’s fair enough, that’s where they’re valued at...That’s their whole reward system. Yes [the special order does get done], but it takes a lot of management effort to tell people that they are going to do it.”

“Setters and leading hands are imbued with this view that the line must not stop, and try as we may we cannot get that out of their thinking. The trouble is we have far too many long serving employees and they know that they have to get 25,000 products off that line this shift and they will do it...They'll believe that they have done a good job and in fact some of the management mechanisms may tell them they're doing a good job. But it might not match the customer service angle and that's what's wrong.”

What happens when they get it right?

“We have a fairly informal management structure. It’s run a bit like a big milk bar.”

Project 3: The Balanced Scorecard – a mechanism to reduce accounting-induced performance anxiety⁴

The Balanced Scorecard (BSC) is a performance measurement innovation designed to specifically counter the adverse effects of managing directly by accounting numbers. The BSC is a much broader performance measurement protocol embracing a wide-ranging set of financial and non-financial metrics that should have a causal link with future profits. At a minimum, the effect should be to dilute the influence of accounting in evaluation, and thus reduce accounting-induced performance anxiety.

BSC captures particularly well the dual role of performance measures, as it is fundamentally designed as a decision-facilitating tool. It provides a dashboard of measures that are basically designed to enhance the information available to managers in decision-making. By targeting leading indicators, managers who seek to make decisions that improve on the BSC metrics should theoretically be implementing strategy effectively and driving future financial performance improvements.

However, there are challenges associated with implementation of a BSC. Some of these challenges relate to the potential for evaluation mechanisms to subvert the good intentions of the BSC. There is also the assumption that the BSC is a neutral management tool that will be used exactly as it is designed to be used: to enhance managerial effectiveness.

The literature has been either silent or equivocal on how this performance measurement innovation interacts with the mechanisms used to evaluate the performance of managers.
Thus, we are potentially back where we started. We have a good decision tool that might not be used if the decisions it signals are not consistent with evaluation mechanisms. What if the manager is expected to use the BSC in decision-making, but her bonus depends on subunit profit? Other researchers have documented that firms initially using a BSC try to reinforce the use of the score card by attaching incentives to performance on the full range of metrics. The problem is that perceptions of validity and reliability of metrics for evaluation and bonuses are somewhat different from the way the same criteria would be applied to information considered relevant for decision-making. There are documented tendencies in practice to over-rely on conventional financial metrics when it comes to evaluation.

This question was addressed in a study⁵ that surveyed 183 profit-centre managers and asked them to identify two sets of performance measures – the measures they consider the most informative for running the business (typically a BSC set of measures) and the measures that are used by the next level of management up the hierarchy to evaluate their performance. It rated the ‘commonality’ between the two sets of measures and assessed the association between ‘commonality’ decisions and profit-centre outcomes.

It found:

Among the managers interviewed, there was a moderate degree of commonality between the measures they considered ‘best’ for running the business and those used to evaluate their performance (approx 62 per cent average commonality).
The weighting on aggregate financial measures in evaluation is significantly greater than their usefulness in running the business (Figure 4). Managers consider disaggregated financial information useful for running the business, such as sales by subunit/product line, costs, cash flows, etc. They consider aggregate financial measures such as profit and ROA less useful. Yet there is a disproportionate reliance on aggregate financial measures in evaluation.
Outcomes improve when firms broaden evaluation protocols to embrace more of the measures that managers consider important. The greater the level of commonality between the two sets of measures, the more the managers actually use the measures that are identified as important in running the business – and by definition, the lower the level of commonality, the less they use these measures. The use of these measures improves the firm’s ability to exploit its capabilities – both existing and future – as managers are using the range of ‘high quality’ measures available to them more effectively. In turn, the cases with higher commonality produce better financial performance outcomes – a result which appears to be a function of greater use of the measures and more strategic responsiveness, not just a direct result of measuring accounting performance and driving improvement on that metric.

Figure 4: Extent of use of performance measures for ‘information’ and ‘evaluation’ Figure 4 Extent of use of performance measures for ‘information’ and ‘evaluation’

Source: Grafton, J., Lillis, A.M. & Widener, S. (2008), op.cit

So, is this the cure? Do we just need to broaden the performance measurement base to incorporate a comprehensive set of non-financial leading indicators that will drive up future profits? The BSC is a performance measurement innovation that has gained significant traction in practice. It has led to a significant shift in performance measurement practice within organisations, in that managers rely less on broad accounting performance measures than they did a decade ago.

However there are still challenges. The BSC is designed primarily to facilitate decisions. Whether or not it manages to do so depends very much on how it links with performance evaluation throughout the firm. There are practical impediments to the complete adoption of the BSC in evaluation, and firms seem reluctant to do it. Many of the issues relate to undue complexity with so many measures, the inescapable ‘softness’ of some measures on a BSC and the fact that they are inherently situation-specific and thus not comparable across subunits. There are many examples of firms reverting to the use of accounting measures in evaluation. To the extent that evaluation protocols remain accounting-focused and disconnected from the array of decision-facilitating measures that managers want to use, performance anxiety will remain and decisions are likely to be accounting-driven.

Concluding comments

This lecture has addressed the issue of accounting-induced performance anxiety from a range of perspectives. Accounting-induced performance anxiety arises from the prominence of accounting in performance measurement practice, and has several consequences. In response to capital market pressures, accounting benchmarks can be shown to have socio-economic consequences as firms change employment patterns when they report accounting losses. There are many ways in which accounting benchmarks can induce short-term decisions that compromise long-term firm value in response to capital market pressures.

Within firms, accounting performance measures create silos, cultivate self-interest and ‘short term-ism’ among managers. It becomes difficult to achieve collaborative ‘firm-wide’ solutions and managers focus on driving up short-term evaluation measures rather than adding long-term value.

There is no apparent cure for capital-market induced performance pressure. Capital markets
will always induce efforts to improve and ‘manage’ accounting performance to convey particular messages about the firm and its prospects. There is much more potential to reduce the adverse impact of accounting-induced decision making within firms, by making adjustments to organisation structure, control systems and performance measurement practices. The BSC is a partial remedy to the dangers of accounting-focused evaluation within firms. However, there are many challenges associated with the reconciliation of the decision-facilitating role of the BSC and the mechanisms used in evaluation. To date there is little evidence of sustainable change in evaluation practice.

¹ Pinnuck, M, & Lillis, A.M. (2007). Profit versus Losses: Does Reporting an Accounting Loss Act as a Heuristic Trigger to Exercise the Abandonment Option and Divest Employees?, The Accounting Review, 82:4, pp. 1031-1053.

² The threshold crossing effects described in the remaining dot points are not evident in Figure 1.

³ Abernethy, M.A. & Lillis, A.M., (1995). The impact of manufacturing flexibility on management control system design, Accounting, Organizations and Society, 20:4, pp. 241-258; Lillis, A.M. (2002). Managing multiple dimensions of manufacturing performance – an exploratory study, Accounting, Organizations and Society, 27:6, pp. 497-529.

⁴ Grafton, J., Lillis, A.M. & Widener, S. (2008). The influence of evaluation mechanisms on the use of decision-facilitating performance measurement information, Working Paper.

⁵ Idem

* I acknowledge helpful comments and/or assistance with the preparation of this inaugural lecture from Margaret Abernethy, Albie Brooks, Marc Costabile, Jennifer Grafton, Richard Lee and Matthew Pinnuck.

Top ^

A condensed version of her Inaugural Lecture delivered at the University of Melbourne on 16 September 2008.

Professor Anne Lillis is Professor of Management Accounting and Deputy Head of the Department of Accounting and Business Information Systems.

Volume 4 NOV 2008