CURRENT COST ACCOUNTING RATIOS AS PREDICTORS OF BUSINESS FAILURE: THE SWEDISH CASE

• Author: Kenth Skogsvik
• Published date: 01 March 1990
• Date: 01 March 1990
• DOI: http://doi.org/10.1111/j.1468-5957.1990.tb00554.x

Journal

of

Business Finance &Accounting,

17(1)

Spring

1990, 0306 686X

$2.50

CURRENT

COST

ACCOUNTING RATIOS

AS

PREDICTORS

OF

BUSINESS FAILURE: THE

SWEDISH

CASE

KENTH

SKOGSVIK*

INTRODUCTION

In various contexts, inflation accounting has been thought

to

be superior to

historical cost accounting. Often inflation accounting has been expected to

provide information which could prove to be helpful in predictions of the

performance

of

a

business company.

For

example, this.expectation was explicitly

stated in Statement of Financial Accounting Standards

No.

33

(p.

4-5):

‘The Board has concluded

that

there is

an

urgent need

for

enterprises

to

provide

information about the effects

on

their activities of general inflation and other price

changes. It believes

that users’

. .

.

ability

to

assess

future

cash

flows

will

be severely

limited until such information

is

included

in

financial

reports.

’

This and similar assertions of inflation accounting have often been founded

on the notion that a common measurement unit is necessary for a ‘fair’

evaluation of

a

business company, and/or some kind

of

deductive reasoning.’

However, an evaluation of the predictive ability of inflation accounting is

essentially an empirical issue.

There have been

a

number of empirical tests of various types of inflation

accounting, pertaining to samples of American or English companies. In the

context

of

predicting business failure, Ketz (1978) and Norton and Smith (1979)

tested general price level accounting, and Mensah (1983) and Keasey and

Watson (1986) evaluated current cost accounting. As a general observation

inflation accounting was not found

to

be superior to historical cost accounting

in these studies. However, looking at estimated costs

of

misclassification, Ketz

(1978) and Mensah (1983) reported

a

weak support of inflation accounting.

Studying the association between various accounting numbers and security

market prices

or

returns, Beaver et al. (1982), Schaefer (1984) and Ohlsen

(1985) did not find any additional explanatory power of current cost accounting

(provided historical cost accounting was available). On the contrary, Bublitz

et al. (1985) did find

a

significant additional explanatory power of current cost

accounting numbers.

*The author is from the Department

of

Accounting and Finance, Stockholm School

of

Economics,

Sweden. The article is based on his Ph.D. dissertation. He wishes

to

thank his advisor, Professor

Sven-Erik Johansson,

for

encouragement and advice.

Also,

he

is grateful for many helpful comments

by Professor Nils Hakansson (University of California, Berkeley), Matti Kinnunen and Eva Widhem.

The computer programming for the project was executed by Matti Kinnunen and Eva Widhem.

Financial support for the research was provided by

Bankforskningsinstitutet

and Carl-Berthel

Nathhorsts Vetenskapliga Stiftelse. (Paper received .July

1987,

revised January

1988)

137

138

SKOCSVIK

In spite of the references above, our knowledge of the information content

and predictive ability of inflation accounting is still fragmentary. Firstly, the

concept of inflation accounting is not unambiguous. There is a choice between

general price level accounting and several different versions

of

current cost

accounting. Secondly, no clear-cut theory for the choice of a specific set of

sufficient

-

from an information point of view

-

accounting numbers is

available. Thirdly,

it

is not clear to what extent the choice of, for example,

multivariate discriminant analysis

(MDA)

might affect a comparison between

current cost and historical cost accounting. In some studies

it

is

most likely

that the assumptions of

MDA

have not been fulfilled

-

cf. Barnes (1987).

Fourthly, few explanations of the weak and rather unstable performance of

inflation accounting have been provided.

The main purpose of this article is to empirically test current cost accounting

(CCA)

ratios regarding the ability to predict business failure for a sample of

Swedish industrial companies.

As

a norm of comparison, prediction results

with historical cost accounting

(HCA)

ratios will be used. In a situation where

published financial statements are mainly based on historical costs, such a norm

could be regarded as obvious. In any way, it can hardly be considered to be

too naive. In several other studies

-

e.g. Beaver (1966), Altman (1968), Blum

(1974), Ohlson (1980) and Zavgren

(1985)

-

HCA ratios have been found

to

be good predictors of business failure.

In a general sense this study provides some additional evidence about the

information content and predictive ability of

CCA.

Like

Mensah (1983), a firm-

specific application of

CCA,

in the sense of Edwards and Bell (1961), is tested.

(Keasey and Watson, 1986, tested a different version of

CCA.)

Obviously, the

sample is unique

-

51

failure and 328 non-failure Swedish companies with

financial statements from mainly the nineteen seventies. Regarding the choice

of

accounting ratios

to

be included in prediction models, a careful investigation

of available information dimensions

in

CCA

and HCA

is

undertaken. Using

probit analysis, prediction models are estimated

1,

2,

3, 4,

5

and 6 years before

failure.

As

the average rate of inflation has varied over time,

it

is possible

to

illustrate a potential association between the rate of inflation and the predictive

ability of

CCA.

The structure

of

the article is as follows. In the next section, the method

of

CCA

is

outlined. The choice of CCA and

HCA

ratios is discussed in the

third section and some data characteristics are presented in the fourth section.

Univariate tests are described and multivariate prediction models are estimated

in the fifth section. Prediction results are reported

in

the sixth section and the

final section comprises some important conclusions and limitations

of

the study.

METHOD

OF

CURRENT COST ACCOUNTING

During the second half of the nineteen seventies, the average (annual) rate

of

inflation was about ten per cent in Sweden. This level implied a considerable