ACCOUNTING FOR PRE-PRODUCTION COSTS IN THE AUSTRALIAN EXTRACTIVE INDUSTRIES

GerhardySenior Lecturer in AccountingSchool of CommerceThe Flinders University of South AustraliaSCHOOL OF COMMERCE RESEARCH PAPER SERIES: 98-2 ISSN: 1441-3906 Abstract This paper examin

Peter G Gerhardy

Senior Lecturer in AccountingSchool of CommerceThe Flinders University of South AustraliaSCHOOL OF COMMERCE

RESEARCH PAPER SERIES: 98-2

ISSN: 1441-3906

Abstract

This paper examines the method used to account for pre-production costs by companies in theAustralian extractive industries Within a costly contracting framework, hypotheses on factorsmotivating the treatment chosen to account for these costs are developed and tested Specifically,the relationship of debt contracting, political sensitivity, income diversification, engineering riskand auditor effect variables to three alternative continuous measures of the accounting treatment aretested While, overall, traditional costly contracting and political sensitivity explanations ofaccounting choice do not appear to hold well, the findings do suggest that the ability to explainobserved variation in accounting treatment is dependent upon the definition of pre-production costsconsidered – particularly whether development costs are included with the early pre-productioncosts relating to explanation and evaluation Further, the results are affected by whetheramortisation is included in the measure of pre-production costs written off Some support is foundfor an effect on the treatment of early stage pre-production costs of the company auditor’sexpressed preference for capitalisation of such costs Evidence is also found to support theproposition that, at the time of the study, those companies in the gold mining sub-sector of theextractive industries were subject to the effects of political sensitivity, related to the impendingremoval of concessional taxation arrangements Some support for a relationship between leverageand the proportion of pre-production costs written off directly to the profit and loss statement is alsoevident

1 Introduction

This paper examines the method used to account for pre-production costs by companies in theAustralian extractive industries Availability of choice for this accounting policy has continueddespite the introduction of Australian accounting standard AAS 7 dealing with the matter in 1976,and companies have and continue to utilise this ability to make a choice in relation to the treatment

of pre-production costs adopted in practice (see for instance, Ryan, Heazlewood and Andrew, 1977,1980; Ryan, Andrew, Gaffikin and Heazlewood, 1990, 1993; Ryan and Heazlewood, 1995, 1997;Gerhardy, 1998) Costly contracting theory is used as the basis for explaining companies’accounting choices Hypotheses relating to factors motivating the treatment chosen to account forpre-production costs in the Australian extractive industries are developed and tested within thisframework

This paper adds to the literature on accounting method choice within the Australianinstitutional framework Its particular contribution to the area is twofold First, it examinescompanies within a particular sector of the economy, namely the mining sector, which has oftenbeen regarded as being subject to differing influences and incentives, particularly in relation to thepolitical sensitivity of firms within the sector (see for example, Bowen, Noreen and Lacey, 1981,

pp 168-9).1 Supporting the importance of taking a single industry approach to accounting choicestudies, Aitken and Loftus (1994, p 2) suggest ‘that future developments in this research willdepend upon the eventual accumulation of research results across a number of separate industries’.The second significant difference between this and many accounting choice studies conducted todate relates to the measurement of accounting treatment adopted The majority of studies in thearea have adopted a dichotomous measure of accounting policy.2 This paper suggests that surveys

of practice which attempt to categorise the choice of treatment for pre-production costs into discreet

1 The possibility of differing influences on sub-sectors of the extractive industries, particularly those in the gold mining and petroleum sub-sectors, is discussed below.

2 Exceptions to this include Aitken and Loftus (1994) who measure the dollar effect of policy choices, and Walker (1988) whose approach has been adopted and modified in this paper.

categories represent a major understatement of the variation found in practice It therefore utilises

an alternative continuous measure of the accounting treatment adopted

In addition to extending the scope of Australian research within the contracting costframework, the study is timely from the point of view of the usefulness of its results to standardsetters Watts and Zimmerman (1986, p 14) argue:

Positive accounting theory is important because it can provide those who must makedecisions on accounting policy (corporate manages, public accountants, loan officers,investors, financial analysts, regulators) with predictions of, and explanations for, theconsequences of their decisions

The Australian standard setters have recently indicated that extractive industries accountingrepresented a high priority project, with the expectation that a discussion paper on the issue would

be prepared in the not too distant future, to be followed by an exposure draft and finally a revisedaccounting standard (AARF & AASB, 1996, p 16) In addition, at its April 1998 meeting theInternational Accounting Standards Committee (IASC) placed on its agenda a new project dealingwith the Extractive Industries A Steering Committee for the project was appointed in July, with thefirst major step in the process to be preparation and issue of a discussion paper by the end of 1998(IASC, 1998) Given the current program to harmonise Australian and International accountingstandards, this move by the IASC suggests that the Australian standard setters will soon commencetheir foreshadowed reconsideration of this country’s extractive industries standards Thus, thisstudy is able to provide further insights into this issue, particularly in relation to the factors that may

be associated with a preference for a particular treatment of pre-production costs

Prior research conducted within the contracting cost framework has indicated threesignificant motivations operating on firms’ accounting method choices, namely managementcompensation contracts, debt constraints and political sensitivity Due to measurement difficultiesthe influence of the first of these factors, the existence of management compensation contracts tied

to earnings, is not investigated in this study The influence of both debt constraints and politicalsensitivity are investigated in detail In addition, the effects on accounting treatment of two factors

hypothesised on the basis of ex ante efficient contracting arguments, namely income diversification

and engineering risk, are investigated Finally, the existence of an auditor effect on accountingmethod, as hypothesised by Deakin (1980), is investigated in the Australian context.

The remainder of this paper is organised as follows The next section briefly outlines thepossible methods of accounting for pre-production costs, and the position adopted by the Australianregulators in relation to this accounting choice Flowing from this, the need for a continuousmeasure of the accounting treatment adopted is explained In Section 3 the formal hypotheses aredeveloped Section 4 provides details of the research methods utilised in the study, with the resultsbeing presented and discussed in Section 5 The final section summarises the major findings anddiscusses the implications of the study

2 Accounting for Pre-production Costs

The fundamental issue distinguishing the different methods of accounting for pre-production costs

in the extractive industries relates to the choice of cost centre to be used in accumulating such costsfor external reporting purposes There exist a large number of possible cost centres, varying greatly

in their size, which could be used to accumulate pre-production costs Lourens and Henderson(1972, p 54) indicate that the cost centre can ‘be as small as an individual mineral deposit, mine,quarry, well, lease or reserve, or it can be as large as several leases, areas, a State, or an entirecountry’ Effectively the choice of a different cost centre leads to the choice of a different method

of accounting for pre-production costs While numerous possible cost centres implies equallynumerous methods of accounting for pre-production costs, Table 1 summarises four of the major

methods available; viz., the full cost, area of interest, successful efforts and expense methods.

[INSERT TABLE 1 ABOUT HERE]

As can be seen from the table, the full cost method is conceptually the opposite of the expensemethod This is by virtue of their adoption of extreme positions as to the cost centre used toaccumulate pre-production costs Both the successful efforts and area of interest methods liebetween the two extremes, with the distinction between the two being less than clear It has forinstance been suggested that the area of interest method is a particular variety of the successful

efforts method, with the cost centre defined as an area of interest (Henderson and Peirson, 1995, p.772; Whittred, Zimmer and Taylor, 1996, p 395) Heazlewood (1987, p 30) on the other handsuggests that the area of interest method lies between the full cost and successful efforts method inview of the fact that the former ‘could result in more expenditure being capitalised than under thesuccessful efforts method in any given geological area’ Davison (1979, pp 22-23) and Wise andWise (1988, p 30) express a similar view Whichever view is accepted, it is suggested that whileTable 1 describes four particular methods of accounting for pre-production costs, this number belies

the wide variety of possible methods which could be used to account for pre-production costs in the

extractive industries, given that the outcome for the two intermediate methods will be dependentupon how the cost centre is defined

In the Australian regulatory context the essential accounting requirements relating to production costs are contained in paragraphs 10-12 of the current version of Australian AccountingStandard AAS 7.3 Paragraph 10 deals with accounting for pre-production costs incurred in the

pre-earlier pre-production stages of exploration and evaluation, and states that:

Costs arising from exploration and evaluation related to an area of interest shall be

written of as incurred, except that they may be carried forward provided that rights to

tenure of the area of interest are current and provided further that at least one of thefollowing conditions is met:

(a) such costs are expected to be recouped through successful development andexploitation of the area of interest, or alternatively, by its sale; and

(b) exploration and evaluation activities in the area of interest have not at reportingdate reached a stage which permits a reasonable assessment of the existence orotherwise of economically recoverable reserves, and active and significant operations

in, or in relation to, the area of interest are continuing (emphasis added)

This paragraph clearly provides a choice of accounting methods for exploration and evaluationcosts While initially it appears to require the expense method, whereby all such costs are writtenoff to the profit and loss statement as incurred, it goes on to add that the costs may be capitalised,subject to one of the stated conditions being met This approach is described as the ‘area of interest

method’ (AAS 7, para 16) Whittred, et al (1996, p 395) suggest that the conditions under which

3 Approved accounting standard AASB 1022 was issued in 1989, giving legislative force for companies to the accounting requirements contained in AAS 7 Since this study refers to accounting practices of mining companies prior to the release of AASB 1022, reference will be confined to the AAS series standard The accounting requirements of the two standards are the same.

this area of interest method may be applied allow ‘the option of deferral even in those cases wherethe feasibility of operations … has not yet been established – all that is required is that rights totenure of the area of interest are current and that the actual and significant operations (not defined)are continuing in that area’ In addition to having an explicit choice between use of the expense andarea of interest methods, the ‘area of interest’ is so broadly defined in geological terms that a widevariety of possible cost centres may be adopted for the purpose of accumulating pre-productioncosts On this point AAS 7 (para 18) admits that during the earlier pre-production stages ‘an areamay be difficult to delimit’ Similarly, Davison and Lourens (1978, p 34) note that in formulatingthe Australian accounting requirements ‘the critical size of the area of interest was, unavoidably,

left undefined in an operational manner, thus permitting a wide variety of options in practice’

(emphasis added) Thus, the available treatments of exploration and evaluation costs are muchbroader than simply a dichotomous choice between two methods, being highly dependent upon howthe area of interest is defined

Exploration and evaluation comprise the initial and generally higher risk pre-productionphases in the extractive industries Accounting for costs incurred in the remaining two pre-

production phases, development and construction, also require consideration Costs incurred in the

latter construction phase generally are in the nature of expenditure on fixed assets and therefore, asrecognised in paragraph 12 of AAS 7, are subject to the accounting requirements of Australian

Accounting Standard AAS 4, Depreciation As such, construction costs are not considered further

in this study In addition, Walker (1988, p 40) suggests that an argument can be put for excludingdevelopment costs when considering accounting for pre-production costs, on the grounds that

‘development typically relates to the preparation of the deposit or field for commercial productionand as such will take place after it has been determined that the exploration has been successful’.This is reflected in the accounting treatment required by AAS 7 (para 11) for such costs, which

‘shall be carried forward to the extent that [they] … are expected to be recouped through successfulexploitation of the area of interest, or alternatively, by its sale’ The choice of methods available for

exploration and evaluation costs is not available when accounting for development costs As such,they may be argued to be different in nature to the costs incurred in the exploration and evaluationphases, and not subject to managerial discretion.

It could however be argued that to exclude development costs would unduly bias the studysince they represent an extra degree of freedom for managers in determining the amount of pre-production costs to carry forward/write off Since, as recognised by Henderson and Peirson (1995,

p 773), ‘Frequently, the boundaries between these phases are unclear as the … phases often occursimultaneously in the same area’, consideration of development costs adds a further dimension tomanagements’ decisions on the accounting treatment to be accorded pre-production costs.4 Thesignificance of this distinction is investigated in this paper by conducting tests which both includeand exclude development costs

The above discussion suggests that specifying the dependent variable in this study asdichotomous, or even as polychotomous, would not accurately represent firms’ accountingtreatments of pre-production costs A refinement of Walker’s (1988) continuous proportionatemeasure is used to represent the accounting treatment of pre-production costs The specification ofthis dependent variable is fully discussed in Section 4

3 Hypothesis Development

Research conducted within the contracting cost framework utilises the contracting and politicalprocesses as the basis for investigating firms’ choices amongst accounting alternatives This sectiondevelops hypotheses relating to the influence of two motivations derived from the framework,namely debt constraints and political sensitivity Also derived from contracting theory, hypothesespredicting engineering risk and income diversification as factors influencing the accounting

4 In addition, a practical argument in favour of considering development costs in this study is the frequent practice of companies whereby all non-construction pre-production costs are included in the balance sheet as a single figure, not allowing in such cases the distinction of development costs from exploration and evaluation costs.

treatment of pre-production costs in the Australian extractive industries are developed Finally,Deakin’s (1980) auditor effect hypothesis is investigated in the Australian regulatory context.

The general hypothesis that the closer is a business to breaching an accounting based debt constraintthe more likely is management to adopt accounting methods that increase profit has been welldocumented in contracting cost theory and related empirical studies (see for instance Watts andZimmerman, 1986, pp 257-259) This strategy is adopted by management in order to loosen suchcovenants and therefore reduce the probability of a breach, and the incurrence of the consequentialcosts

The most commonly used proxy for closeness to debt covenant default in prior studies hasbeen a leverage ratio.5 Whittred and Zimmer (1986) and Stokes and Tay (1988) provide evidence

of the use of various types of restrictive covenants in trust deeds supporting listed public debt issues

by Australian companies Both studies indicate that most commonly such covenants are specified

in terms of liabilities to total tangible assets Whittred and Zimmer (1986, p 27) suggest that

exclusion of intangibles means that for a number of accounting choices, including inter alia., the

treatment of pre-production costs, to a large extent ‘the choice of capitalizing or expensing cannotaffect a firm’s proximity to a leverage constraint’ This would seem to suggest that the general hypothesis outlined above would not hold in the current study Despite this argument, two factorssuggest that investigation in this study of the relationship between leverage and the choice ofaccounting treatment for pre-production costs is warranted First, both of the surveys mentionedinvestigated only public debt agreements The current study suggests that public debt is seldomused in the Australian extractive industries, with only one company included in the study havingsuch debt outstanding It is possible that covenants in private debt agreements differ from those forpublic debt issues In addition, Whittred and Zimmer’s (1986) survey includes only one companywith a stock exchange industry classification indicating involvement in the extractive industries,

5 As recognised by Press and Weintrop (1990, p 65) such proxies for closeness to breach of covenants are used because of the high cost of gaining access to actual contracts.

with none of the companies in Stokes and Tay’s (1988) survey being involved in the pre-productionstages of the extractive industries Thus, to draw the conclusion that pre-production costs areroutinely precluded when specifying leverage constraints in debt agreements may be somewhatpremature In addition, Press and Weintrop (1990) found evidence in the US context of anassociation between firms’ level of leverage and closeness to debt constraints.

Based on the above, and recalling that capitalisation of pre-production costs will have theeffect of reducing the probability of breaching a debt covenant specified in terms of the debt toequity ratio, the following hypothesis is tested:

H1: The ratio of debt to equity will be positively correlated with capitalisation of

pre-production costs.

A further ratio often specified in debt covenants is the level of interest coverage Asobserved by Bowen, Noreen and Lacey (1981, p 167), ‘Some debt agreements prohibit the firmfrom issuing new debt unless a minimum prescribed ratio of income to interest charges ismaintained’ Whittred and Zimmer (1986, p 25) provide evidence of the use of such interestcoverage constraints in Australia, although they were found in only 33 per cent of the debenturetrust deeds they examined The potential difficulty of leverage ratios being unaffected by themethod of treatment of pre-production costs if intangible assets are excluded in the specification ofthe constraint does not apply to constraints specified in terms of interest coverage This thereforesuggests a possible observable relationship between the interest coverage ratio and the accountingtreatment of pre-production costs in the Australian extractive industries Specifically, capitalisation

of pre-production costs in preference to writing them off will have the effect of increasing the level

of interest coverage above what it would be under the alternative, thereby reducing the probabilitythat the firm will violate such a constraint Thus the lower the level of coverage the more likely isthe firm to capitalise such costs The following hypothesis is tested:

H2: The interest coverage ratio will be positively correlated with writing off or

expensing of pre-production costs.

3.2 Political Sensitivity

Contracting cost theory establishes a link between the political sensitivity of firms and theiraccounting method choices, hypothesising that the more politically sensitive is a firm the morelikely is management to choose income reducing accounting methods Traditionally studiesinvestigating political sensitivity have adopted firm size, measured in various ways, as the proxy forpolitical sensitivity.6

Malmquist (1990, p 181) provides an additional ex ante argument as to why firm size may

be related to the accounting method choice of US oil and gas companies He argues that whilechoosing to capitalise pre-production costs rather than write them off has the effect of decreasingvariance in reported profits, the effectiveness of this strategy reduces as firm size increases Heargues that larger firms will have a greater number of drilling projects, which leads to a ‘portfolioeffect’, by diversifying away variance due to drilling risk.7 Thus, for such firms the necessity tocapitalise pre-production costs to reduce earnings variance is diminished Conversely, the potentialexists for smaller firms to gain much in the way of variance reduction by adopting a policy ofcapitalisation, since they are less likely to have the number of separate drilling projects required to

gain variance reduction from the ‘portfolio effect’ This ex ante efficiency argument implies, as does the ex post political sensitivity argument outlined previously, that firm size will be correlated

positively with the expensing of pre-production costs

Revenue and total assets have been the two most frequently adopted proxies for firm size inprior accounting choice studies In their study Deegan and Hallam (1991, p 11) adopt what they

consider to be an alternative and preferable measure of firm size, viz., net profit after tax and before

extraordinary items They base their preference for this measure on Watts and Zimmerman’s

(1986, p 239) view that a firm’s accounting earnings are ‘a better proxy for the negative/positivecorporate wealth transfers [than] firm size (total sales or assets)’ Thus, substituting a net profit

6 Lim (1996) provides a comprehensive review of the literature relating to the political sensitivity/costs hypothesis.

7 The concept of drilling or engineering risk and its likely effect on accounting treatment of pre-production costs

in the extractive industries is discussed further in Section 3.3 below.

measure for a measure of size such as revenue or total assets may provide a better proxy forpolitical sensitivity by providing a more direct measure of political costs The following hypothesis

is tested:

H3: Net profit after tax and before extraordinary items will be positively correlated

with writing off or expensing of pre-production costs.

Watts and Zimmerman (1986, p 235) suggest that ‘The most direct way to transfercorporate assets [from firms to the government] is via the tax system, and therefore, income taxes

are one component of political costs borne by firms’ (emphasis in original) However, it is not a complete measure of political costs since, inter alia., it does not take into account the possibility

that wealth transfers through the tax system may be offset by various measures, such as the granting

of subsidies Companies in the Australian extractive industries have been subject to particularconcessional taxation treatment, the main concession being the extensive capital expendituredeductions for the cost of exploration, prospecting and mining operations (Income Tax AssessmentAct 1936, Divisions 10, 10AA and 10AAA) Further, income from gold mining and sales was, until

1 January 1991, exempt from tax Also, as noted by Walker (1988, p 23), companies in theAustralian extractive industries, and in particular those involved in the petroleum industry, mayhave been subject to political threat from the late 1970s to the mid 1980s by a proposed resourcerent tax This tax was finally introduced in 1987 (Petroleum Resource Rent Tax Act 1987) It maytherefore be argued that the accounting treatment adopted during 1986, the year considered in thecurrent study, would reflect managers’ motivations in relation to the economic consequences to beincurred due to impending introduction of the new tax, outlined in detail at that time in thePetroleum Resource Rent Tax Bill 1986 Since the tax was to be levied on ‘economic profits’ (Balland Bowers, 1984, in Walker, 1988, p 23), managers of those firms potentially affected by the newtax may have been expected to adopt income decreasing accounting methods

Following from the above discussion, it is argued that the effective tax rate may be used as ameasure of the political sensitivity of companies in the study Companies with higher effective taxrates, it is argued, are subject to greater political costs, and so would have an incentive to reduce

their political sensitivity by writing off pre-production costs rather than capitalising them Thefollowing hypothesis is tested:

H4: The ratio of tax expense (net of deferred taxes) to net profit before tax and

extraordinary items will be positively correlated with writing off or expensing of pre-production costs.

In addition, the preceding discussion suggests that the imposition of political costs may fallunevenly upon firms in different sub-sectors of the extractive industries For instance, companiesinvolved in gold mining and exploration, and those in the petroleum sector, would appear to bedistinguishable from other sectors of the extractive industries due to extant or potential differentialtax treatments To the extent that the political sensitivity variables discussed above do not fullycapture the effects on pre-production cost method choice arising due to industry sub-sectormembership, this may be controlled for by the use of categorical/dummy variables.8 The discussionsuggests that managers of firms involved in the petroleum sector of the extractive industries would

be more likely to adopt a method of accounting for pre-production costs which is incomedecreasing, in order to reduce the political visibility associated with the impending imposition of theresource rent tax Similarly, the incentive for managers of firms in the gold sector to adopt suchincome decreasing methods is likely to be particularly strong Such firms would wish to protecttheir privileged position with respect to the concessional tax treatment received by reducing theirpolitical visibility The following hypotheses, which capture these industry sub-sector effects, aretested:

H5: Involvement in the petroleum sector of the extractive industries will be

positively correlated with writing off or expensing of pre-production costs H6: Involvement in the gold sector of the extractive industries will be positively

correlated with writing off or expensing of pre-production costs.

Studies by Deakin (1979) and Lilien and Pastina (1982) suggest the possible influence of factorssuch as aggressiveness in exploration and exploratory risk upon the method chosen to account for

8 See the full specification of the model in Section 4.

pre-production costs in the US oil and gas industry The former study found no significantrelationship between the surrogates used for aggressiveness in exploration and the accountingmethods adopted Lilien and Pastina (1982) hypothesise that firms undertaking extensiveexploration activities incur proportionately more unsuccessful activities involving relatively largeamounts of pre-production costs Such firms are argued to have a greater incentive to capitalisethese costs in preference to writing them off, in order to avoid significant deterioration in theirreported income and asset balances Further, such action would have the effect of decreasing thevariability of reported income, and therefore possibly the perceived level of risk Their studysupports this hypothesis by finding exploratory risk, measured as the proportion of dry wells to totalwells, to be significantly related to choice of an income maximising accounting policy Morerecently Malmquist (1990) argued that the relative magnitudes of two types of risk, namely drilling(or engineering) risk and product market risk, determine the degree of variation in a firm’s earnings.

He defines engineering risk (p 181) as ‘the probability of failure in exploration activities’ This

risk will be greater for firms with activities concentrated in exploration vis-à-vis those with their

resources concentrated in production.9 Firms in the former group seek to report higher income byadopting an accounting treatment that maximises the pre-production costs capitalised.10 Malmquist(1990) tests the proposition that the proportion of a firm’s resources devoted to drilling andexploration is positively related to the likelihood of choosing an accounting method whichcapitalises the maximum amount of pre-production costs (in the US context, the full cost method)

9 Malmquist (1990) argues that depending upon their age, size and experience firms will develop a comparative advantage in either the exploration or production phases of operations, and therefore direct their resources in order

to exploit this comparative advantage.

10 Firms with resources concentrated in production will report higher income by adopting an accounting method which writes off pre-production costs earlier This keeps the amounts of such costs to be amortised against production revenue to a minimum, thereby maximising the amount of profit reported.

He adopts as a measure of the importance of exploration ‘the ratio of the firm’s world-wideexploration costs to the year-end market value of its equity’ In this study the measure used to testthe influence of engineering risk uses the book rather than market value of equity The followinghypothesis is tested:11

H7: The ratio of exploration costs incurred to the book value of total equity will be

positively correlated with capitalisation of pre-production costs.

In her study of the 1974 accounting practices of companies in the Australian extractiveindustries, Walker (1988, p 30) suggests an alternative measure which could be used to capture theincentive for managers to minimise negative income effects She proposes that the more diversifiedare the operations of a firm the more likely they are to expense pre-production costs, since theexistence of diversified sources of income would increase the capacity of the firm to bear the writeoff without producing undue negative earnings effects The results of Walker’s (1988, pp 55-57)study do not support the diversification hypothesis However, this may, in part, be due to the timing

of the study, which related to 1974 Undiversified firms at this time may not have felt it necessary

to mitigate negative earnings effects by capitalising pre-production costs rather than expensing themsince, as suggested by Walker (1988, p 57), ‘the revenues generated from the production of oil, gasand minerals, are of such a magnitude that the income effects of a write off … of pre-productioncosts will be mitigated’ Such an incentive may be related, at least in part, to the state of theoil/mineral market at the time, which is known to change markedly over time For instance, theyear 1986, of concern in the current study, saw substantial falls in the price of petroleum products,which may increase the incentive to diversify revenue sources It would therefore seem worthwhile

to retest the hypothesis in the context of the current study However, measuring a company’s

11 Malmquist’s hypothesis relating product market risk to the propensity to choose a method which capitalises less pre-production costs is not tested in this study due to the difficulty of constructing an appropriate surrogate with accessible data Production figures for Australian firms in the extractive industries are not available on a firm-by- firm basis and, due to their commercially sensitive nature, are unlikely to be disclosed upon request This omission may not unduly bias the model given that firms with high engineering risk will have relatively low product market

risk, and vice versa Use of a single variable to represent engineering risk may therefore be sufficient to capture the

effects of risk on the accounting treatment adopted.

diversification opportunities based on the number of lines of business does not take into account therelative significance of these lines in earning revenue for the company A finer measure of thediversification effect, which provides a better indication of the firm’s capacity to bear the write off

of pre-production costs than simply the number of additional lines of business, is the proportion oftotal revenue earned from sources other than the extractive industries The following hypothesis istested:

H8: The ratio of revenue earned from extra lines of business, other than those

involving extractive industries operations, to total revenue will be positively correlated with writing off or expensing of pre-production costs.

12 See for instance Watts and Zimmerman (1981), Chow (1982) and DeAngelo (1982).

methods’, it is not possible to discern from the study whether this is, in fact, the appropriateexplanation for the observed association Deakin’s (1980) hypothesis is investigated in theAustralian context in the current study Specifically, it is suggested that auditor’s expressedpreference for either the expensing or capitalisation of pre-production costs will be associated withthe treatment of such costs by their clients in the Australian extractive industries The followinghypothesis is tested:13

H9: Auditor preference for writing off vis-à-vis capitalising pre-production costs will

be positively correlated with their clients’ accounting treatment of such costs.

4 Research Method

The population of concern in this study is all companies listed on the mining board of the AustralianStock Exchange during 1986 As indicated in the previous section, the differential tax treatment ofincome from gold mining operations and the influence upon the petroleum industry of the proposal

to introduce a resource rent tax can both be investigated by choosing 1986 as the year ofinvestigation Also, previous extensive studies of accounting for pre-production costs in Australiahave concentrated around the time of the original issue of AAS 7, in 1976 The persistence ofvariation in practice since release of the standard, and the likelihood of an impending review of thestandard, both discussed in Section 1, suggests choice of a more recent period may provideadditional useful insights

The specific companies included in the study were determined by the source of much of thedata used in the study, namely the AGSM Annual Report File All companies with 1986 annualreports on the file and which were listed on the mining board of the stock exchange were identified

As indicated in Table 2, of the 123 such companies found on the file, 104 were finally included inthe study The reasons for exclusion of the other 19 are indicated in the Table.14, 15

13 The methods used to operationalise this hypothesis are detailed in the next section.

14 A list of the companies in the study is available from the author.

15 Utilisation of the AGSM Annual Report File as the data source for the study may affect the external validity of the results obtained By design, the file, being is based on the top 500 Australian companies by market

[INSERT TABLE 2 ABOUT HERE]

The dependent variable in this study is the accounting treatment of pre-production costs Section 2

of the paper outlined the rationale for using a continuous proportionate measure of the dependentvariable, rather than the dichotomous, or even polychotomous classification of accounting treatmentfrequently adopted in accounting choice studies.16 As argued previously, firms have a great deal oflatitude in the amount of pre-production costs to be expensed and/or capitalised in any year In thisstudy the continuous measure used by Walker (1988, p 40), the ratio of all pre-production costsexpensed or written off for the year to the opening balance of pre-production costs carried forward

at the beginning of the year, has been modified by adopting a different denominator in the ratio.17

In this research the measure used is the ratio of all pre-production costs expensed or written off for

the year to the closing balance of pre-production costs carried forward plus pre-production costs

expensed or written off during the year The denominator therefore represents the total amount ofpre-production costs which theoretically could have been written off during the year Thus, thedependent variable can be described as the proportion of the total amount of pre-production costsavailable to be written off which was actually written off during the year

Initially the dependent variable is defined to include all non-construction pre-productioncost; that is, exploration, evaluation and development costs (WRITEOFF1) However since, asoutlined in Section 2, there are arguments supporting the exclusion of development costs from thedefinition, additional tests are performed using the more restricted definition (WRITEOFF2),

capitalisation, contains proportionately more larger companies than would be included in the general population of extractive industries companies.

16 A detailed accounting policy review of the 104 companies included in the study was undertaken to determine the accounting method for pre-production costs adopted by each company While the vast majority adopted the

‘area of interest method’ required by the standard, many of these companies developed their own interpretations of the appropriate way to implement the method in practice This review therefore confirmed the belief expressed previously, that in practice it would be an oversimplification to say that the area of interest method represents a single standard method of accounting for pre-production costs Detailed results of this policy review are not reported here due to space limitations.

17 As noted by Walker (1988, p 40) a proportionate measure is required since the absolute amount of such costs expensed in a year (the numerator) will be closely related to firm size.

utilising the reduced sample of companies for which it is possible to distinguish exploration andevaluation costs from development costs.

A further issue relating to the definition of the dependent variable is the extent of production costs written off to include in the numerator Specifically, in addition to those pre-production costs written off directly, should it also include amounts of pre-production costsamortised as a cost of production? Walker (1988, p 55) suggests that ‘a finer measure’ of thedependent variable, which excludes those costs expensed which represent amortisation of costscarried forward, may provide a different result On the other hand, amortisation of pre-productioncosts represents an accounting technology over which management has some degree of discretion,and is an integral part of the overall treatment of pre-production costs and the decision to expense orcapitalise such costs It can therefore be argued that it is appropriate to include any amortisation ofpre-production costs in the numerator of the dependent variable In this study the initial dependentvariable (WRITEOFF1) adopts the broader measure of pre-production costs written off, withadditional tests, using the restricted measure of the numerator (WRITEOFF3), conducted todetermine the sensitivity of the results to this distinction The primary measure of the accountingtreatment of pre-production costs adopted in this study is WRITEOFF1, the definition of which isprovided in Table 3 In addition the two more restricted measures, labelled WRITEOFF2 andWRITEOFF3, are defined and are utilised for comparison throughout As for all variables in thisstudy, unless otherwise indicated, the data used to calculate their values are taken from the financialstatements and notes contained in the companies’ 1986 annual reports

pre-[INSERT TABLE 3 ABOUT HERE]

Two proxies are used in this study to capture the effects of debt constraints on managements’treatment of pre-production costs, one based on leverage constraints and one on interest coverageconstraints These are defined in Table 4 as LEVERAGE and INTCOVER respectively The

predicted signs for these variables are indicated in line with the hypothesised relationship of thevariables to the dependent variable, as detailed in hypotheses H1 and H2.

[INSERT TABLE 4 ABOUT HERE]

While both of these variables relate to the use of debt in a company’s capital structure, theyrepresent different aspects of the impact of debt constraints, one related to balance sheet amountsand one to profit and loss statement amounts Both LEVERAGE and INTCOVER are to someextent interdependent with the dependent variable measures outlined previously In the case of the

former variable, ceteris paribus, the lower is the amount of pre-production costs written off the

greater will be the company’s equity Thus, the lower will be the calculated value of LEVERAGE.Given that the dependent variable is continuous in nature, rather than involving a discrete choicebetween well defined accounting methods, it is not possible to sensibly adjust for this dependency;however, as pointed out by Zimmer (1986, p 47), the effect of undoing the difference in accountingtreatment in such cases would be to increase the likelihood of leverage being related to theaccounting treatment adopted That is, the unadjusted measure used in this study errs on the side ofconservatism by making a finding of significance more difficult to achieve A similar dependency

is present in the definition of INTCOVER Again the unadjusted measure can be seen asconservative in that it makes it more difficult to achieve a finding of the hypothesised relationship

It was suggested in Section 3 that net profit may provide a better measure of politicalsensitivity than the size measures traditionally used in accounting choice studies The independentvariable PROFIT, defined in Table 4, is therefore used as the prime proxy for political costs in thisstudy.18 The amount of PROFIT is also affected by the treatment of pre-production costs adopted

by the company, and again the unadjusted measure used is conservative in that it makes a finding ofsignificance less likely

Another political sensitivity variable utilised in this study is TAXRATE, argued in Section 3

to represent a partial measure of political costs imposed on firms TAXRATE is dependent on the

18 For completeness all tests involving the PROFIT variable were re-run using in its place first the book value of total assets (ASSETS) and second total revenue (REVENUE) Any significant change in results is reported in the relevant section.

accounting treatment of pre-production costs adopted to the extent that net profit will be lower thegreater is the amount of such costs written off This implies that not undoing this dependencymakes a finding of a significant positive relationship more likely Thus, any results finding such arelationship need to be interpreted with this in mind Forty-three companies in the study reported anet loss before tax Of these, eight reported positive income tax expense For such firmsTAXRATE was set to one The remaining 35 loss companies reported either zero or negative taxexpense for the year For these TAXRATE was set to zero.19

The final two political costs variables defined in Table 4 are PETROL and GOLD Theseare both (0,1) dummy variables relating to industry involvement, since participation in thepetroleum and gold sectors of the extractive industries are hypothesised to be related to thetreatment of pre-production costs (H5 and H6) Three separate sources of data were used todetermine companies’ involvement, or otherwise, in the petroleum and gold sectors First, the thenAASE industry classification codes for each company in the sample was checked A code of 015indicated ‘gold’ involvement, while any code from 031 to 034 inclusive indicated involvement in

‘oil and gas’ While this represented a first attempt at determining industry involvement, thepublished AASE code alone does not provide a sufficient guarantee of correct and completeclassification for a number of reasons First, there may be errors in the published AASE code, orthe published code may be out of date if companies have recently changed or expanded theirinvolvement in particular sectors Also generic codes, such as 019 ‘Mineral Exploration’ and 243

‘Diversified Resources’, do not specify the particular type of mineral/petroleum product involved

For this reason, information on corporate activities contained in Jobson’s Mining Yearbook

1987-1988 (30th edition), which includes information on companies up to 1986, was used to crosscheck

the industry involvement of sample companies As a final check, information contained in theannual report was used to check the classification provided by the two prior sources Based on thisprocess industry involvement was determined and values of zero or one were assigned to the twodummy variables

19 This follows a similar procedure adopted by Wong (1988, p 157).

Table 4 provides definitions of the explanatory variables relating to engineering risk andincome diversification ENGNRISK is designed to measure the relative importance of explorationexpenditure, as discussed in relation to hypothesis H7.20 The variable EXTRALNS measures theimportance of extra lines of business, other than extractive industries activities, as contributors tothe companies’ revenue stream, calculated from segment information provided in the companies’annual reports.

The final two variables contained in Table 4 are definitions of the two (0,1) dummyvariables used to model the effect of auditors’ preference for capitalisation of pre-production costs

vis-à-vis immediate expensing of such costs as incurred While the definitions are self-explanatory,

the method of determining auditor preferences requires explanation This study is concerned with

1986 accounting practices As such, inquiries of audit firms, by mail or telephone survey, someyears later are unlikely to provide an accurate indication of their preferences at the time in question

A secondary source of this information was therefore sought One source of the views ofindividuals or organisations on accounting policies is the submissions made to the AARF when an

exposure draft is released and comments on its contents are invited In relation to Statement of

Accounting Standards AAS 7, the most recent opportunity for submissions to be made on the issue

of the treatment of pre-production costs occurred with the issue in January 1989 of Release

416/Exposure Draft 47, which, inter alia., invited comments on the appropriateness of the standard

permitting a choice of methods While it is recognised that the timing of these submissions to theAARF is some three years after the year examined in this study, they represent the most reliable andcontemporaneous source of information on preferences available

Eleven of the 95 submissions made were classified as by large accounting firms, includingall of the then ‘Big-8’ firms Since all of these firms acted as auditor for at least one companyincluded in this study, their submissions were examined to determine whether a preference for a

20 As noted in Table 4, when the dependent variable is measured as WRITEOFF2, which excludes development costs, an alternative measure of engineering risk, ENGNRISK2, which also excludes such costs from its numerator,

is used.

particular treatment to account for pre-production costs was expressed.21 Preferences were divided

into three categories, viz., those favouring the immediate write off of such costs, those favouring

allowing their capitalisation (possibly subject to certain conditions being met) and those which wereneutral on the matter Submissions which did not mention the matter were classified as neutral, as

were those which favoured the status quo of allowing a choice between the two treatments None

of the firm’s submissions indicated a preference for the immediate expensing of pre-productioncosts In addition to those firms making submissions to the AARF, 19 companies included in thestudy employed accounting firms which did not make such submissions Since such firms did notfeel compelled to express their views on the issue when given the opportunity they were regarded asbeing neutral on the matter of the preferred treatment of pre-production costs, and were classified assuch Table 5 indicates the number of sample companies for which each of the accounting firmsacted as auditor, and indicates whether each firm was classified as preferring capitalisation of pre-production costs (C) or as being neutral on the issue (N).22

[INSERT TABLE 5 ABOUT HERE]

It should be noted that Table 5 lists 12 accounting firms, rather than 11, in the category of firmsmaking submissions to the AARF The reason for this anomaly is the merger of the firms KMGHungerfords and Peat Marwick Mitchell & Company between 1986 and 1989 Thus, while the twoseparate firms were in existence and acted as auditor for sample companies in 1986, by 1989 whensubmissions were made to the AARF they had become a single firm, KPMG Peat MarwickHungerfords Since determination of the separate firm’s preferred treatment of pre-production costswas not possible, it was decided to classify both the individual firms as having the preference

expressed in 1989 by the merged firm, viz., capitalisation of such costs It is recognised that this

classification may misrepresent what might have been the separate preferences of the two firms,

21 Thirty-eight of the submissions were classified under the heading ‘Small Practices’, but none of these firms acted as auditor to any company in this study.

22 Since no accounting firm acting as an auditor for a sample company is classified as having a preference for the immediate expensing of pre-production costs, the variable AUDITEXP becomes redundant All the relevant accounting firms are classified as either preferring capitalisation or being neutral on the matter Thus a dichotomous, rather than trichotomous, classification results, with the single (0,1) dummy variable AUDITCAP being sufficient to fully describe this situation.