Thesis (PhD)--University of Stellenbosch, 2005. / ENGLISH ABSTRACT: Several mixed industry prediction models about failure have been reasonably
successful in differentiating between successful companies and those that have
failed. The challenge, however, is to venture into the grey area in between and to
identify companies, which are financially unstable, at an early stage. Early
identification enables management to intervene timeously in an attempt to prevent
failure. Failure is defined as either liquidation, delisting, suspension of listing or a
substantial change in structure.
The grey area focused on in this study is overtrading. Overtrading is triggered by the
company growing at too high a rate relative to its specific structure. Cash is
necessary to fund expansion, whether for an increase in inventories, credit sales or
new non-current assets. If the company does not generate enough cash to fund this
expansion, it has to be financed through external sources. The longer the period of
growth and the higher the growth rate, the more the cash requirements.
From the theoretical model underlying overtrading, it was found that:
• the higher the growth in sales,
• the smaller the profit margin, and
• the higher the net current assets in proportion to total assets,
the lower the cash flow from operating activities before dividends were paid (CFO).
Any company ought to generate enough cash from its daily activities in order to
maintain the existing level of business, to repay loans, to replace assets and to pay
dividends. If the internal generation of cash is insufficient to finance these activities,
existing cash resources will be consumed, unproductive non-current assets will be
sold and possibly also some of the productive non-current assets. The outcome for
such a company is a business combination or liquidation. Due to the fact that cash
plays such a big role in failure, cash flow variables constitute the majority of the
independent variables used in the development of the failure prediction models.
The overtrading ratio was developed as a measurement tool to quantify overtrading.
As long as the company generates a positive CFO, it is not so much at risk as a
company that does not succeed in generating a positive CFO. Therefore, a negative
CFO for a three-year period was decided on as the norm for identifying possible financial difficulty. A company is involved in overtrading if the sum of CFO for three
years less the sum of the adjusted profit for the three years, divided by the absolute
value of the sum of the adjusted profit for the three years equals -1 or smaller in the
case of a company with a cumulative profit for the three years; and smaller than
nought in the case of a company with a cumulative loss for the three years.
South African industrial companies listed for at least three years during the period
1974 to 2003, were identified. From a total of 6 662 cumulative three-year periods,
944 overtrading years were identified. Failure occurred in 212 out of 526 companies
involved in overtrading between January 1974 and August 1989. 120 out of 199
companies involved in overtrading between September 1989 and November 1995
failed, while 90 out of 127 companies involved in overtrading, failed between
December 1995 and June 2000. By June 2005 it was already evident that 49 out of
92 companies involved in overtrading between July 2000 and December 2003, had
already failed.
Companies involved in overtrading, may survive artificially for lengthy periods with
the support of providers of capital. It can therefore be expected that failure prediction
models will not achieve a better accuracy rate than achieved by probabilities. Six
failure prediction models utilising classification tree algorithms were developed. Using
data from two periods, two different models were developed; one for growth and
recession phases of the economy, the other without distinction between economic
phases. The first period was September 1989 to June 2000, the other December
1995 to June 2000. June 2000 was chosen as the cut-off, since a period of five years
after an overtrading year was necessary to follow-up whether the company had
failed. Each universe was split in two – the learning sample, more or less 60%, and
the test sample, more or less 40%. The models were developed from the learning
sample and the test sample was used as substantiation of the results of the
developed model.
The total classification accuracy of the three best models, one for the growth-phase,
one for the recession-phase and one mixed economy model, is respectively 72,99%,
96,67% and 80,26% and the classification accuracy for the failed companies 75,29%,
100% and 85,19% respectively. The total prediction accuracy of the three models is
respectively 69,23%, 80,95% and 72,55%, and that of the failed companies 73,68%,
86,67% and 83,33%. The accuracy of all the models was found to be higher than what the accuracy would have been if all the companies involved in overtrading were
merely classified as having failed.
From the results of the different tests, it seems that Ver3, the growth in sales from
year 1 to year 3, is probably the most important independent variable in the
classification between failed and non-failed overtrading years. This corroborates the
theory underlying overtrading that indicates that a high sales growth puts a company
at risk for cash flow problems.
Companies where the cash flow problems develop because of an increase in current
assets will be intercepted by the overtrading ratio. Companies where cash flow
problems develop due to replacement of non-current assets, will not necessarily be
intercepted by the overtrading ratio as CFO that is used in the overtrading ratio does
not allow for replacement of non-current assets. It is therefore necessary to adjust
CFO to a free cash flow CFO.
Depreciation is used as an alternative for replacement investment since disclosure of
replacement investment is not required. Depreciation is theoretically the fraction of
the value of an asset lost during the year; this value needs to be replaced. By
subtracting the depreciation for the year from CFO, this amount will be more
representative of the cash position of the company after considering all the normal
transactions in order to sustain the business.
After all the adjustments for a free cash flow, six models were developed for the
different periods and economic phases. The accuracy of these models were better
than what the accuracy would have been if overtrading years were merely classified
as failed. Implementing these models would therefore improve specificity.
From the tests performed, Ver3 and KVB3/TB (the cumulative CFO for the three
years over total assets) seem to be the most important independent variables in the
classification between failed and non-failed when considering free cash flow. This is
informative as KVB3:TB represents a fictional amount, as if the company spent an
amount equal to depreciation on replacement investment. / AFRIKAANSE OPSOMMING: Heelwat gemengde industrie-mislukkingvoorspellingsmodelle was al redelik
suksesvol in die onderskeid tussen mislukte en suksesvolle maatskappye. Die
uitdaging is egter om die grys area tussen dié uiterstes te betree en ’n maatskappy
wat finansieel onstabiel is, vroegtydig te identifiseer. Vroegtydige identifikasie stel die
bestuur in staat om betyds in te gryp en mislukking te voorkom. Mislukking word as
likwidasie, òf denotering, òf opskorting van notering, òf ’n wesenlike
struktuurverandering, gedefinieer.
Die grys area waarop die fokus in hierdie studie val, is oorbedryf. Oorbedryf word
veroorsaak deurdat die maatskappy teen ’n te hoë koers relatief tot sy spesifieke
struktuur groei. Kontant is nodig om uitbreiding, hetsy ’n toename in voorraad,
kredietverkope of nuwe nie-bedryfsbates, te finansier. Indien die besigheid nie
genoeg kontant genereer om hierdie uitbreiding te finansier nie, moet dit vanuit
eksterne bronne gefinansier word. Hoe langer die tydperk van groei voortduur en hoe
hoër die groeikoers is, hoe meer is die kontantbehoeftes.
Uit die teoretiese model wat onderliggend is aan oorbedryf, is bevind dat:
• hoe hoër die groei in verkope,
• hoe kleiner die winspersentasie, en
• hoe hoër die bedryfskapitaal in verhouding tot die totale bates;
hoe laer is die kontantvloei uit bedryfsaktiwiteite voor dividende betaal (KVB).
Enige besigheid behoort genoeg kontant uit dag-tot-dag-aktiwiteite te genereer ten
einde die bestaande vlak van besigheid vol te hou, lenings terug te betaal, bates te
vervang en dividende te betaal. Indien die interne kontantgenerering onvoldoende is
om hierdie aktiwiteite te finansier, sal bestaande kontantbronne uitgewis word,
onproduktiewe nie-bedryfsbates sal verkoop word en moontlik ook van die
produktiewe nie-bedryfsbates. Die uiteinde vir so ’n maatskappy is ’n besigheidsoorname
of ’n likwidasie. Aangesien kontantvloei so ’n groot rol in mislukking speel,
is kontantvloeiveranderlikes die meerderheid van die onafhanklike veranderlikes wat
in die ontwikkeling van die mislukkingvoorspellingsmodelle gebruik is.
Die oorbedryfsratio is as ’n maatstaf ontwikkel om oorbedryf te meet. Solank as wat
’n maatskappy ’n positiewe KVB genereer, is hy nie so riskant soos ’n maatskappy wat nie daarin kan slaag om positiewe kontant te genereer nie. Daarom is ’n
negatiewe kumulatiewe KVB vir ’n drie-jaar-tydperk as die norm gestel om moontlike
finansiële nood te identifiseer. ’n Maatskappy is besig met oorbedryf indien die som
van die KVB vir drie jaar minus die som van die aangepaste wins vir die drie jaar,
gedeel deur die absolute waarde van die som van die aangepaste wins vir die drie
jaar, gelyk aan of kleiner as -1 is, in die geval van ’n maatskappy wat ’n kumulatiewe
wins vir die drie jaar het; en kleiner as nul in die geval van ’n maatskappy wat ’n
kumulatiewe verlies vir die drie jaar het.
Alle Suid-Afrikaanse genoteerde industriële maatskappye wat vir ten minste drie jaar
gedurende die tydperk 1974 tot 2003 genoteer was, is geïdentifiseer. Uit ’n totaal van
6 662 kumulatiewe drie-jaar-tydperke was daar 944 oorbedryfsjare. Vanaf Januarie
1974 tot Augustus 1989 was daar 526 oorbedryfsjare, waarvan 212 misluk het. Vanaf
September 1989 tot November 1995 was daar 199 oorbedryfsjare, waarvan 120
misluk het en vanaf Desember 1995 tot Junie 2000 het 90 van 127 oorbedryfsjare
misluk. Teen Junie 2005 was dit reeds bekend dat 49 van die 92 oorbedryfsjare
tussen Julie 2000 en Desember 2003 misluk het.
Oorbedryfsmaatskappye bly soms vir uitgebreide tydperke kunsmatig voortbestaan,
indien die verskaffers van kapitaal hulle dra. Die verwagting was dus dat
mislukkingvoorspellingsmodelle nie ’n beter akkuraatheid sou behaal as wat
waarskynlikhede sou bepaal nie. Ses mislukkingvoorspellingsmodelle is met behulp
van die klassifikasieboomalgoritme ontwikkel. Een model elk vir die groeifase en die
resessie-fase van die ekonomie en een model sonder onderskeid van die
ekonomiese fase is met die gebruikmaking van twee tydperke se data ontwikkel. Die
eerste tydperk was September 1989 tot Junie 2000 en die ander Desember 1995 tot
Junie 2000. Junie 2000 is as die afsnypunt gebruik aangesien ’n vyf-jaaropvolgtydperk
na ’n oorbedryfsjaar nodig is om vas te stel of die maatskappy misluk
het. Elke universum is in twee verdeel – die leersteekproef, ongeveer 60%, en die
toetssteekproef, ongeveer 40%. Die modelle is uit die leersteekproef afgelei en die
toetssteekproef is gebruik as bevestiging van die resultate van die afgeleide model.
Die totale klassifikasie-akkuraatheid van die drie beste modelle, een vir die groeifase,
een vir die resessie-fase en een gemengde ekonomiemodel, is onderskeidelik
72,99%, 96,67% en 80,26% en dié vir die mislukte maatskappye 75,29%, 100% en
85,19%. Die totale voorspellingsakkuraatheid van die drie modelle is onderskeidelik 69,23%, 80,95% en 72,55% en dié van die mislukte maatskappye 73,68%, 86,67%
en 83,33%. Al die modelle se akkuraatheid is meer as wat die akkuraatheid sou wees
indien al die oorbedryfsjare bloot as misluk geklassifiseer sou wees.
Uit die resultate van verskeie toetse blyk dit dat Ver3, die groei in verkope vanaf jaar
1 tot jaar 3, waarskynlik die belangrikste onafhanklike veranderlike in die onderskeid
tussen mislukte en nie-mislukte oorbedryfsjare is. Dit ondersteun die teorie wat
onderliggend is aan oorbedryf, wat aandui dat ’n hoë groei in verkope ’n maatskappy
op risiko plaas vir kontantvloeiprobleme.
Maatskappye waar die kontantvloeiprobleme vanweë ’n uitbreiding in bedryfskapitaal
ontstaan, word deur die oorbedryfsratio onderskep. Die maatskappye waar die
kontantvloeiprobleme eerder uit die vervanging van nie-bedryfsbates sal voortvloei,
sal nie noodwendig deur die oorbedryfsratio ondervang word nie, aangesien die KVB
wat in die oorbedryfsratio gebruik word, nie voorsiening maak vir vervangende
investering nie. Dit is daarom nodig om KVB tot ’n vrye kontantvloei-KVB aan te pas.
Waardevermindering word as ’n alternatief vir vervangende investering gebruik,
aangesien vervangende investering nie ’n verpligte openbaarmakingsvereiste is nie.
Waardevermindering is teoreties dié deel van die bate wat gedurende die jaar
opgebruik is en wat vervang behoort te word. Deur dus die jaarlikse
waardevermindering van KVB af te trek, is hierdie syfer meer verteenwoordigend van
die maatskappy se kontantposisie nadat alle normale transaksies om die besigheid in
stand te hou, in ag geneem is.
Na die aanpassings vir vrye kontantvloei is weer ses modelle uit die verskillende
tydperke en vir die verskillende ekonomiese fases ontwikkel. Al die modelle se
akkuraatheid is beter as wat die akkuraatheid sou wees indien al die oorbedryfsjare
bloot as misluk geklassifiseer sou wees.
Ver3 en KVB3:TB (die kumulatiewe KVB vir die drie jaar tot die totale bates) blyk uit
die toetse die belangrikste onafhanklike veranderlikes te wees in die onderskeid
tussen misluk en nie-misluk wanneer vrye kontantvloei in ag geneem word. Dit is
insiggewend aangesien KVB3:TB ’n fiktiewe syfer verteenwoordig, sou die
maatskappy ’n bedrag wat gelyk is aan waardevermindering, aan vervangende
investering bestee het.
| Identifer | oai:union.ndltd.org:netd.ac.za/oai:union.ndltd.org:sun/oai:scholar.sun.ac.za:10019.1/16521 |
| Date | 12 1900 |
| Creators | Steyn, Barbara Wilhelmina |
| Contributors | Hamman, W.D., Smit, E. vd M., University of Stellenbosch. Faculty of Economic and Management Sciences. Dept. of Business Management. |
| Publisher | Stellenbosch : University of Stellenbosch |
| Source Sets | South African National ETD Portal |
| Language | Afrikaans |
| Detected Language | English |
| Type | Thesis |
| Format | xxviii, 438 leaves : ill. |
| Rights | University of Stellenbosch |
Page generated in 0.0035 seconds