You are on page 1of 5

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/297312888

Sunk costs

Article  in  Appita Journal · April 2010

CITATIONS READS

0 177

1 author:

Geoff Covey
Covey Consulting
71 PUBLICATIONS   268 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

chemicals from biomass View project

All content following this page was uploaded by Geoff Covey on 13 January 2017.

The user has requested enhancement of the downloaded file.


Feature

Sunk costs
This paper, by GEOFF COVEY1 was presented at the 53rd Appita Annual Conference and Exhibition,
Melbourne, April 2009

Sunk costs are expenditure that has already been made with text books and references, but experience shows that the con-
some objective, but which cannot be recovered if that objective cepts are poorly understood and many inappropriate decisions
is abandoned. By their nature they relate to a specific objective are made as a result.
(e.g. a non-refundable deposit on a particular piece of land or
equipment) such that a decision not to proceed with the purchase WHEN SUNK COSTS BECOME IMPORTANT
will result in the forfeiture of the deposit. In some cases only part
The first thing to understand is that sunk costs usually only come
of the expenditure is lost – e.g. in the purchase of equipment, it into consideration after the decision to undertake a project and after
might be possible to transfer part of the deposit for the purchase expenditure has been made. An exception to this should be the
of more expensive equipment. recognition before starting work that such exercises as research and
An everyday example of a sunk cost is a ‘no changes permitted’ feasibility studies will be sunk costs whatever the outcome.
airline ticket. If one no longer needs or wants to take the trip for Sunk costs usually become important because something has
which the ticket was booked, then the money is lost (sunk). Even gone wrong; the project has not worked as expected, market con-
if one decides to take the trip anyway, but without any purpose, the ditions have changed or the company’s objectives have been
money should be regarded as sunk because neither the original changed. In these cases the attractive return on investment that
objective, nor any reasonable alternative objective has been met. was predicted a priori will no longer be met and a decision must
A slightly more complex situation arises with ‘partially flexi- be made as to what is to be done with the assets of the project
ble’ fares. Here the ticket is not refundable, but can be credited under current circumstances.
towards another ticket with the same airline of equal or greater When this happens, the costs expended in the past are irrele-
value, provided that a re-booking fee is also paid. In this case, if vant.
the original flight is no longer required, in which case the entire All that is relevant is future costs and income.
purchase cost is lost (sunk), or the re-booking fee is paid in Obviously, a company that keeps making investments and
which case only an amount equivalent to the re-booking fee can then writing off large chunks of expenditure and proceeding
be regarded as the equivalent of sunk. purely on the basis of future cash flow will not stay in business
The concepts and consequences of sunk costs are generally for very long. The objective must be to salvage the maximum
quite simple, but many bad business decisions are made because value when things have not turned out as expected.
of improper consideration or treatment of sunk costs. (In the examples that follow, simple return on original invest-
Properly, sunk costs must be treated as irretrievable and part ment type comparisons are made. A discounted cash flow
of the past. Any decision involving circumstances in which costs approach is, of course, more appropriate but the differences are
have already been sunk should be made on the basis of ignoring relatively small for most of the examples and the objective is to
costs which have already been sunk and considering only keep the analysis as simple as possible. Use of DCF methods
future expenditure. would not alter the principles of the issues to be considered).
Failure to treat the issue of sunk costs is sometimes due to a
poor understanding of the economic analysis of the case, but it A hole in the ground
often has a strong subjective (or emotional) component. There is Based on seismic surveys and test drillings, a mining company
a tendency to ignore proper economic analysis and instead be has invested AUD 3 million in sinking a deep mine shaft to
swayed by one of the following: exploit a valuable mineral deposit. Unfortunately, when the shaft
• ‘Honouring sunk costs’ (not wanting to spoil the ship for a is completed it is found that the actual deposit is of poorer grade
ha’penny-worth of tar). This is a tendency to stick with the and worse mineralisation than survey work had indicated and it
original plan even though the circumstances have changed is not worth proceeding further.
(1,2) (it is interesting to note that much of the literature on A geological research group needs a deep hole to position
this aspect of the issue has been published in journals related geo-monitoring equipment, and offers the company AUD
to psychology or philosophy rather than economics). 200,000 for the mine shaft.
• Abandonment (not throwing good money after bad). This is As this represents less than 7% of the cost of sinking the shaft,
a tendency to over-compensate and to completely abandon the company rejects the offer.
the investment rather than to consider ways of making a par- This is a fallacy, and an example of confusing cost with value
tial recovery of costs (3). (4), as well as a failure to understand sunk costs (in this case lit-
This paper will attempt to show how sunk costs should be erally). The hole may have cost AUD 3 million to dig, but with
treated objectively and to provide simple examples of how fail- no resources at the bottom its value is zero. Further, it is not an
ure to do this can result in poor financial decisions. It is stressed asset that can be used in a different way by the company (mine
that most of the concepts presented here can be found in standard shafts cannot be relocated to a more promising site).
The company should accept the offer of AUD 200,000
1Chairman because, irrespective of the cost of sinking the shaft, the offer is
Covey Consulting Pty. Ltd. 1st Floor, 660 High St. Kew VIC 3102 AUD 200,000 more than the value of the hole.
98 Appita Journal Vol 63 No 2
Feature

Let it rot? re-sell them for much more than half their purchase price.
Suppose the circumstances are as before, except that fabrica-
A farmer has cultivated, planted the seed and applied fertiliser
tion of the reactor vessel is nearly complete, and the fabricator
and insecticides at a total cost of AUD 30,000. All that remains
still offers you 50% of the material value for it.
is to spend another AUD 10,000 in harvesting the crop.
You should still not accept. Scrap platinum is usually worth
Before he planted, the anticipated selling price of the crop
much more than 50% of merchant stocks (a search of the web
was AUD 60,000 and he expected a profit of AUD 20,000 (this
will show dealers offering about 80% of spot price for industri-
analysis ignores opportunity costs for the field).
al platinum scrap).
However, the yield in his area is lower than normal, while
The plant also had a similar configuration reactor made from
there is a glut of produce elsewhere, and the anticipated selling
low grade stainless steel. Fabrication is almost complete when
price is now only AUD 25,000. Should he harvest or allow the
the project is cancelled. Again the fabricator has offered 50% of
crop to rot?
the material value.
Had he known the real value of the harvested crop in advance,
This time you probably should take the offer. The reactor is
the farmer would not have planted. However circumstances have
unlikely to be useful to anyone else and so the fabrication costs
changed and costs incurred to date should be ignored. His real
should be regarded as sunk. The value of the reactor is only its
choice is do nothing and lose no further money, or to spend an
scrap value and for low grade stainless steel, 50% is probably a
additional AUD 10,000 to generate AUD 25,000.
good price.
He should harvest and only be down AUD 15,000 on the year
rather than let the crop rot and be down AUD 30,000.
The obsolete filter
This is a case where good money should be thrown after bad
to reduce overall loss. Your mill purchased a new leaf filter for AUD 400,000 two
years ago. Another manufacturer has brought out a much
Pointless advertising improved alternative that only costs AUD 240,000 to buy and
has an operating cost of AUD 18,000 a year (compared with
You have spent AUD 500,000 advance advertising a new,
AUD 30,000 for the present filter).
patented product for which you have sole agency to import.
Should you replace the filter?
Now the company which owned the patents and was to have
This particular problem can be looked at in several ways.
manufactured the product has gone bankrupt and stopped all
Firstly on the basis of conventional profit analysis (figures in AUD):
operations.
Existing filter New filter
What can you do about it? Probably nothing.
Purchase cost 400,000 240,000
What is the value of the advertising? Probably nothing.
Life when new 10 10
Depreciation 40,000 24,000
The non-refundable deposit
Ann op cost 30,000 18,000
You have agreed to buy some equipment for AUD 100,000 and Years in use 2 0
have paid a non-refundable deposit of AUD 30,000.
You have now been offered the same piece of equipment by Total annual cost 70,000 42,000
another supplier for AUD 75,000. What should you do?
This can be looked at in two ways, and both show that you Simple profit considerations suggest that the filter should be
should complete the original purchase. replaced. Similarly, cash flow analysis also suggests replacement.
To have the equipment, you can either complete the original However, this is not the correct way to analyse the situation.
purchase for an additional outlay of AUD 70,000 (the deposit is The existing filter has already been purchased and the deprecia-
a sunk cost) or you can switch to the low supplier for an outlay tion charges applied to it are not true costs. To compare the total
of AUD 75,000. annual costs for the filter that has already been purchased with
An alternative way of looking at this which avoids the con- the new one, depreciation should be included for the new filter
cept of sunk costs in the normal sense is to consider that the cost but not for the existing filter because the purchase price of the
of going with the new supplier is AUD 105,000, being the actu- existing filter is a sunk cost, but the purchase of the alternative
al purchase price plus the cost of forfeiting the deposit. The cost filter lies in the future and is not a sunk cost.
of going with the original supplier remains at AUD 100,000. On this basis, the annual cost of the existing filter is reduced
to AUD 30,000, while the annual cost of the new filter is
The cancelled reactor unchanged at AUD 42,000.
You are planning to build a new manufacturing plant that The analysis above is somewhat oversimplified as it ignores
includes an unusual high temperature reactor using chlorine as various scrap value and tax benefit issues. A more complete
a reactant. For this purpose you have purchased 500kg of plat- analysis is as follows:
inum sheet and delivered it to the fabricator. The filter has been operating for two years; therefore if it is
Because of the economic down-turn, the project is suddenly being depreciated at 10%/year straight line, the book value of
cancelled. the filter would be AUD 320,000.
The fabricator offers you 50% of the purchase price for the Suppose it can be sold second hand for AUD 50,000, then the
platinum sheet (platinum prices have been stable since you loss on disposal would be AUD 270,000.
bought the material). Should you take it? At a tax rate of 30%, this results in a tax credit of AUD
NO. This is not a sunk cost situation. Raw materials have been 81,000. This credit and the sale proceeds can be applied to the
purchased, but are still in original form. It should be possible to purchase of the new filter (bringing the cost of changing over
April – May 2010 99
Feature

down to AUD 109,000) and the rest of the loss on disposal is ing with the sunk cost of the original investment.
written off (it is a sunk cost). The following table ignores sunk costs and looks at cash-
In this case the comparison becomes (figures in AUD): flows for the seven year remaining life of the existing plant
Existing filter New filter Cashflow(AUD million) Keep old Displace
Net purchase cost 109,000 Year 1 -50
Life when new 10 10 3 13
Depreciation ignore 10,900 Year 2 3 13
Ann op cost 30,000 18,000 Year 3 3 13
Years in use 2 0 Year 4 3 13
Year 5 3 13
Total annual cost 30,000 28,900 Year 6 3 13
Year 7 3 13
The depreciation used above is not the value for tax purposes, Total 21 41
but it gives the most favourable ‘proper’ case for purchasing the
new filter. There is now a slight benefit in replacing the filter. Clearly, the concern about depreciation is irrelevant and the exist-
Suppose instead, the original filter has nor been delivered yet ing facility should be displaced by one using the new technology.
and the order can be cancelled at a cost of 25% of the purchase In fact, the table understates the benefits of going to the new
price (figures in AUD): technology, as there will also be the tax advantage of an immedi-
Go with existing Buy new plus ate write-off of the old facility, and the displacing plant will still
cancellation fee have three years of life after the seven years considered above.
Purchase cost 400,000 240,000
Cancellation fee 100,000 Good money after bad
Cost incl canc fee 340,000
Life when new 10 10 A company spent AUD 10 million on a project with an antici-
Depreciation 40,000 34,000 pated profit of AUD 1 million/year. However, the plant can only
Ann op cost 30,000 18,000 be operated at low throughput and is generating a loss of AUD
Years in use 2 0 1.5 million/year.
The problem has been traced to one section of the plant which
Total annual cost 70,000 52,000 is inappropriate for the actual (rather than anticipated) operating
conditions. This problem can be resolve by the expenditure of a
In this case it does make more sense to cancel the existing further AUD 1 million after which the plant will generate a prof-
order and buy the new model filter. it of AUD 0.7 million/year.
Note in the above analysis, the cancellation fee has been What should the company do?
added to the cost of the new filter as it is not incurred if the orig- Standard analysis would be:
inal order proceeds. It is also assumed that for comparison pur- If proceed:
poses (but not for tax) the cancellation fee is written off over the Total cost AUD 'million 11
life of the equipment. Profit AUD 'million/year 0.7
Return on investment 6.36%
Displace or replace
The previous example is one where errors in decisions are very This return on investment is too low, spending more would
commonly made (including errors not to change equipment be throwing good money after bad, and the project should be
even though this is the proper option). Many examples can be abandoned.
found in the literature e.g. Humphreys (5) who refers to the This is wrong.
problem as ‘Displacement Vs Replacement’ (Replacement If the project is abandoned, the AUD 10 million will be lost
being substituting for a worn out item, and Displacement being anyway, so it should be left out of the consideration (it is a sunk
substituting for an item that is still functional, but not econom- cost). This also means that depreciation of the past AUD 10mil-
ic). Levine (6) gives a qualitative discussion of a real (but dis- lion should also be neglected. So the current cash flow is -AUD
guised) case where a poor decision was made because some 0.5million – still too low.
equipment had not been fully written off. Consider the case of performing the remedial work and treat-
A similar case, but analysed quantitatively follows: ing sunk costs correctly:
A company spent AUD 50 million three years ago building a This is now a project where only AUD 1million is to be
production facility. There are still seven years of depreciation to spent. Only depreciation on this additional investment is con-
be written off and the plant has negligible scrap or resale value. sidered. Therefore the cash flow from the upgraded plant is
The technology used in the facility was made obsolete almost AUD 1.8 million/year and the profit is AUD 1.7 million/year.
immediately and it is now making a loss of AUD 2million/year, The investment outcome can be summarised thus:
with a positive cash flow of AUD 3million/year. Capital cost 1
It is possible to adopt the technology being used by competi- Depreciation 0.1 On new investment only
tors by building a new plant costing AUD 50million which Cash flow 1.8
would generate a profit of AUD 8million/year (and a cash flow Profit 1.7
of AUD 13 million/year). Return on investment 170% Pretty good!
Although the old plant is not fully depreciated, the important
point here is that depreciation is not a real cost, and we are deal- Obviously the upgrade should be undertaken.
100 Appita Journal Vol 63 No 2
Feature

A tale of two miners uneconomic and stayed there for four years. BHP tried unsuc-
cessfully to sell the smelter. By late 2003 they had mines and a
Two interesting cases arise from BHP Billiton.
smelter that were uneconomic to operate and which no one
In November 2008 BHP Billiton abandoned its year-long
wanted to purchase; irrespective of what BHP had paid for
take-over bid for Rio Tinto. In doing so it wrote off some AUD
3-600 million invested in the deal. them, their value was zero. Further, continuing care and mainte-
This is a powerful example of a sunk cost and of a response nance was costing the company money and the longer a plant
to changing circumstances. When BHP Billiton launched its bid remains in this mode the more difficult and expensive it is to
the commodities markets were booming and both companies restart it.
appeared to be in strong positions. The commodity prices At the time it was made, the decision to regard the purchase cost
slumped and the values of shares in both companies followed as sunk and to stop throwing good money after bad was correct.
them down. Rio Tinto faced the problem of servicing large This example shows several important points about sunk
amounts of debt (which has resulted in it selling off assets). costs and irreversible decisions. The original purchase of
BHP Billiton decided that the offer was no longer in the inter- Magma Copper was based on past copper prices and reasonable
ests of its share holders and withdrew its offer. expectations of future prices. Shortly after the purchase the mar-
This was a clear case where circumstances had changed dra- ket unexpectedly entered a prolonged slump. BHP attempted to
matically since the original decision was made and the compa- sell the assets for less than they had paid – i.e. recognised that
ny decided to abandon the money already spent rather than to the original purchase price was a sunk cost and that this cost had
‘throw good money after bad’. no bearing on the current value of the assets. At the time of the
Probably the company was also cautious after its experiences decision to demolish the value of the assets were effectively nil
with changing commodity prices shortly after it purchased and BHP stopped further expenditure.
Magma Copper. The decision to demolish was one to formally sink the invest-
BHP purchased Magma Copper in 1996 when copper prices ment. It is unfortunate for BHP that copper prices then rose after
were at about USD 2600/t and rising. it was too late for them to benefit.
Almost immediately copper prices steadied and then began a This illustrates the point that the special treatment of sunk
severe decline. Through the first half of 1999 prices were in the costs normally only matter if circumstances change after the
USD 1400-1500/t and BHP stopped production at the San investment decision was made. It is also quite possible that after
Miguel smelter in Arizona and put the plant onto ‘care and the decision to dispose of an asset that there will be further
maintenance’. unforseen changes in circumstances.
For the next four years copper prices remained fairly steady
at this low price, and in October 2003 BHP announced the per- World scale, commodity machines only
manent closure of its Arizona copper operations and the demo- A final example may be seen from the attitude of the paper
lition of the smelter. industry towards old machines.
There was an immediate rise in copper prices, at first thought The last 30 years has seen the local industry (and the indus-
just to be a market reaction to one large producer withdrawing. tries in many other developed countries) shut down small
But the rise continued and by February 2004, copper prices machines because they no longer fit (often after the machine
were higher than when BHP purchased Magma. From April was switched from making low tonnage but profitable products
2006 to September 2008 copper traded at between USD 5700/t to commodity grades for which it was not suited).
and USD 8400/t (usually towards the upper end of this range. There was an opportunity that was rarely taken or even consid-
After this the price of copper plummeted in response to the ered to recognise that the machine was now available at zero cost
world financial crisis, but in early February 2009, it was still and switch in opposite direction to high margin specialty grades.
trading at above USD 3300. The price of copper and dates of
BHP’s actions are shown in Figure 1. CONCLUSIONS
With hindsight, the decision to demolish the smelter was a
bad one and made at just the wrong time. Had BHP waited Two important lessons should come from this paper:
another three months they would probably have restarted it Firstly, certain types of expenditure, such as research and
instead. However, in terms of the treatment of costs as sunk, at advertising, are sunk costs from the moment that they are made,
the time when the final decision was made, the action was prob- and if the project does not proceed these costs can rarely be
ably correct. BHP had spent AUD 3.2 billion to buy what looked recovered.
like a sound asset. Shortly after the purchase the gains in the Secondly, when the real world differs from that assumed
market reversed and copper fell to a price that made production when the decision to implement the project was made, past costs
should be ignored when planning future strategies about the pro-
$10,000 ject assets.
BHP purchases Magma
$9,000
$8,000 REFERENCES
$7,000 (1) Knox, R.E. and Inkster, J.A. ‘Postdecision dissonance at post time’ Journal
Plant on care
$6,000 and maintain of Personality and Social Psychology 8:319-323 (1968)
$5,000 (2) Bass, R ‘Sunk Costs’ Department of Philosophy, Costal Carolina
$4,000 University, www.geocities.com/amosapient/sunkost.pdf
(3) Bruno, S and De Lellis, A ‘The economics of ex ante co-ordination’
$3,000
Workshop on ‘Systems of Innovation’, EU, CNRS-LATAPSES,
$2,000 Shut and Dipartimento di Scienze Economiche dell'Università di Bologna, Bologna
$1,000 demolish 5-6 October 1992
$0 (4) Shore, D. and Covey, G. ‘Balancing Cost and Value’ Proc. 60th Appita
1993 1995 1998 2001 2004 2006 2009 2012 Conference, Melbourne, 2006, p119.
(5) Humphreys, K.K. Jellen’s cost and optimisation engineering 3rd Edn.
Fig. 1 Copper prices since 1993, and BHP’s actions with McGraw-Hill, 199, p196 et seq.
Magma Copper (Copper prices from Indexmundi) (6) Levine, H.A. ‘Sunk costs don’t matter’ PM World Today XI(1) (Jan 2007).

April – May 2010 101


View publication stats

You might also like