Supply Risk Management at
Unilever
Managing Spend at Risk
01/2009-5563
This case was written by INSEAD Professors Paul Kleindorfer and Enver Ycesan, in cooperation with the Supply
Management Lead Team of Unilever Corporation. The issues raised here are purely for educational purposes and
are not intended to illustrate either effective or ineffective management of an administrative situation
Copyright 2009 INSEAD
N.B.: TO ORDER COPIES OF INSEAD CASES, SEE DETAILS ON BACK COVER. COPIES MAY NOT BE MADE WITHOUT PERMISSION.
Pro
babi
lity
Upper limit
Annual Spend 08
Sept 07
March 08
Pro
babi
lity
Upper limit
Annual Spend 08
Sept 07
March 08
609-006-1
ecch the case for learningDistributed by ecch, UK and USA North America Rest of the worldwww.ecch.com t +1 781 239 5884 t +44 (0)1234 750903All rights reserved f +1 781 239 5885 f +44 (0)1234 751125Printed in UK and USA e [email protected] e [email protected]
Copyright 2009 INSEAD 1 01/2009-5563
The Supply Management Leadership Team (SMLT) at Unilever brought together the heads of key procurement areas such as oils and fats, plastics, cocoa, and milk powder at Unilever. The SMLT was preparing for a special meeting in Zurich in a month to discuss possible changes to its risk management tactics and strategy. The mood among team members was unusually somber. In the past few months, commodity prices had defied all the laws of gravity by jumping to historically high levels. Uwe Schulte, Vice President of Global Supply Management, had asked his team to come up with innovative approaches to commodity procurement that would provide a better understanding, and perhaps some mitigation, of Unilevers exposure to the increased volatility of market prices. The presentations of the risk management proposals by the plastics team was the top agenda item for the meeting, but it was understood that every one of Unilevers major commodity purchases would be subject to a similar review over the next few months. Given the developments of the past few months, the focus was on understanding and managing the risk of large swings in procurement expenditures. How to do this without sacrificing buying performance and a dependable physical supply to Unilevers manufacturing facilities would be the centerpiece of the SMLT meeting and discussion.
Unilever
Unilever was a global giant in food and personal care products. Operating in 150 countries with 206,000 employees, its turnover was 39.7 billion euros in 2005. Figure 1 shows the distribution of its activities across different categories and different regions.
Home Care
18%
Spreads
11%
Personal Care
26%
Savory &
Dressings
21%
Ice Cream and
Frozen Food
16%
Beverages
8%
Europe41%
Americas33%
Asia/Africa26%
Figure 1: Distribution of Unilevers Activities Around the World
Unilever had global brands such as Lipton, Knorr, Lux and Omo that were top brands in their categories, as well as locally strong brands such as Hellmanns, Birds Eye, Carte dOr and Axe. Twelve of these brands achieved annual sales volumes of more than 1 billion euros each.
609-006-1
Copyright 2009 INSEAD 2 01/2009-5563
In a press release1, Patrick Cescau, Unilevers CEO, laid out an ambitious agenda:
We now need to drive harder to build a winning portfolio by extending our leadership positions and our presence in high growth spaces. At the same time, we are improving our consumer marketing and customer development to deliver outstanding execution. Bringing all this together as One Unilever will ensure that we capitalize on both our local roots and global scale. This strategy will enable us to grow ahead of our markets with sustainable margin improvement. I am confident that this will lead to sustainable underlying sales growth of 3-5% and an operating margin in excess of 15% by 2010.
SMLT
Unilever was organized upon three pillars: categories, regions, and functions. The Supply Management (SM) organization, which was responsible for global procurement, was part of the Supply Chain Management function. SCMs mission was to build one supply network from shelf to supplier, which leveraged Unilevers scale and delivered competitive solutions to customers and consumers. The organizational structure is depicted in Figure 2. SMLT consisted of the heads of the regions and of key spend categories such as chemicals, food ingredients, and packaging.
Figure 2: The Supply Management Organization
1 3 August 2006
SMLT
SCLT Sponsor Greg Polcer
John Rice
UEx Sponsor
Globally responsible for SM on behalf of UL
SCLT
FOODS Category
FINANCE Function
HPC Category
Reg I ons
EUROPE
AMERICAS
ASIA
AMET
David Beauchamp/
Guenther Buck (O&F)
Marco Gonalves
Umesh Shah
Krish Maharaj
6 global teams
7 global teams
6 global teams
3 global CoPs
PACKNET Uwe G Schulte
VP Packaging Network VP Global SM
NPI Peter Pick
VP NPI
CHEMNET INGNET Henk Sijbring
VP Chemicals Network/ HPC Category Contact
Jan-Jelle vd Meer Dir Ingredients Network Foods
609-006-1
Copyright 2009 INSEAD 3 01/2009-5563
SM had several guiding principles. With respect to the supply base, it strived to achieve competitive costs, be the innovation partner of choice and a professional business partner. With respect to the downstream supply chain, SM aimed at delivering superior customer service. Under Uwe Schultes leadership, the SMLT also considered efficient risk management part of its mandate, and the new initiative on managing market risk for major commodity groups was a central aspect of this element of SMLTs responsibilities.
Commodity Characteristics
SM purchased a wide range of commodities, including food ingredients (such as sugar, powdered milk, cocoa, wheat, and various oils and fats); chemical products (such as lab, caustic soda, and alcohol sulphates); packaging materials (such as plastics, aluminum, and corrugated cardboard); as well as energy (such as electricity and natural gas). The total annual spend under SMs responsibility was several billion euros but the profit consequences of a dependable supply at predictable prices clearly went well beyond the direct impact of its annual spend. Indeed, the sharp rise (Figure 3) and increased volatility (Figure 4) in commodity prices had drawn a lot of attention to SMs activities and further heightened the pressure to improve profitability following a few disappointing quarters for all the major companies in the food industry and the resulting stock market reactions. In the context of this dual pressure, SMLT decided to evaluate alternative risk mitigation strategies for its sizeable commodity procurement business.
Figure 3: Year-on-Year Change in Commodity Procurement (constant volumes)
-800
-600
-400
-200
0
200
400
600
1999
2000
2001
2002
2003
2004
2005
2006m
n
609-006-1
Copyright 2009 INSEAD 4 01/2009-5563
Figure 4: Increased Volatility in Commodity Prices
Risk Management Options
As a pilot project, SMLT decided to begin with plastics to understand what benefits risk management innovations could provide. As with most major commodities, parallel financial markets were well developed for plastic resins and the recent price volatility had created heightened awareness of the need to limit exposure. Table 1 summarizes some recent market movements in key spend items for selected commodities under SMs responsibility.
Table 1: Market Characteristics of Selected Commodities
-40%
-20%
0%
20%
40%
60%
80%
1999 2000 2001 2002 2003 2004 2005 2006
Tallow
Dairy EU
Dairy NA
Alkoxylated
Resins
LAB/LAS
O&F
Cluster
Maximum historic
annual price volatility
Parallel Financial market? Tools available
Natural Gas EU 30% Limited OTC SwapsDairy ingredients EU 25% NoCocoa 30% LME, OTC Futures + OptionsDairy NA 30% CME, limited Futures+OptionsPAS/AE/LES 25% NoPlastics 25% LME (limited) + OTC Futures + SwapsLAB/LAS 60% OTC SwapsOils & Fats 25% CBOT + OTC Futures
609-006-1
Copyright 2009 INSEAD 5 01/2009-5563
Plastics
Plastics for containers come in many sizes and grades, but the underlying chemical components remain the same, with the basic chemicals (referred to as resins) being high-density polyethylene (HDPE), polypropylene (PP) and polyethylene terephthalate (PET). While SM did not purchase these resins directly, its suppliers of plastic containers did, and they passed on the cost of these resins to SM via the price of the plastic containers it purchased. SM was a major consumer of plastic bottles and lids to package Unilevers products, spending around 400 million euros on plastics in 2006. Price volatility peaked at 25% within the year.
In spite of a complex supply chain, as depicted in Figure 5, plastics were highly commoditized. As a key commodity, the global over-the-counter plastics trade was quite transparent with well developed parallel financial markets since the price of both PP and HDPE was highly correlated with the price of their raw materials, namely crude oil and natural gas. This suggested that hedging instruments for plastics spend could be either directly in HDPE, PP or PET, or through positions in the underlying raw materials of crude oil and natural gas.
SM purchased bottles and lids from dedicated suppliers in each of its major sales regions. Bottle blowers closely followed crude oil prices and resulting resin prices before adding their own margins on the product. Due to product volumes, shipping was also quite expensive. SMs strategy for this category was to achieve price stability. Unfortunately, market forecasts were far from accurate. This, in turn, triggered SMs interest in pursuing parallel risk management opportunities, both as a possible source of risk hedging as well as to improve the quality of the information underlying SMs decisions regarding sourcing and contracting for plastics.
Figure 5: Overlapping HDPE, PP, and PET Supply Chains
PP and HDPE Chains Overlap
Ethylene
Aromatics PTA/DMT
Naphtha (1t) PP
Others: L(L)DPE/PVC/MEG/PET
(polyester) fiber/PETGasoline additives +
others
Ethane HDPE20%Natural
Gas
Crude oil
Transport/ fuel
Propylene
80%
Methane95%
95%
Reds show competing markets for feedstock use
Various markets outside the resin market affect resin prices. Crude oil (energy and transportation) is the most prominent affecting both PP and HDPE.
60%
0.32t
0.16t
0.10t
609-006-1
Copyright 2009 INSEAD 6 01/2009-5563
Figure 6 shows the structure and complexity of SMs plastics hedging problem. For a specific region, the problem begins with aggregate end product demand for different bottles and containers, as determined by Unilevers market forecast. These, in turn, imply ex ante demand for different resins (HDPE, PP, PET) in the periods t = 1, , T (think of these as quarters). The basic question confronting the SMLT in terms of hedging strategies was whether to take positions (i.e., buy swaps and other derivative instruments from brokers or directly on the London LIFFE or other exchanges) in crude oil (an indirect hedge) or in the resins themselves. At this juncture SMLT was not interested in changing its physical sourcing of plastics from its direct plastics manufacturers, but only in understanding the cost and value of various hedging strategies.
Figure 6: Demand for Plastic Bottles
As an example, Table 2 summarizes demand and mean prices for three demand regions (NE, SE and West) in the North American market for HDPE for the four quarters of 2006 (as forecast at the beginning of 2006). The standard deviation of HDPE demand in each quarter was expected to be about 10% and the standard deviation of the price was expected to be about 25% of mean price, with strong correlation of prices across regional suppliers. During 2006, based on futures contracts trading on 1/1/06, the average price for Brent crude oil was expected to be $65.32 per barrel (with a standard quarterly deviation of $6.17). For the same period, the correlation between crude oil prices and that of resins was around 0.65.
D1t D2t Dmt .
HDPEt
PETt
PPt
Crude
Oil Price
Market
Traded Resins
Final Product
Demands
Correlations Correlations
Resin
Demands
609-006-1
Copyright 2009 INSEAD 7 01/2009-5563
Quarter 1 Quarter 2 Quarter 3 Quarter 4
Demand-NE 15,557 16,159 14,532 16,858
Demand-SE 16,550 16,734 14,985 15,485
Demand-We 19,150 17,810 18,283 18,895
Price-NE 1,211 1,304 1,405 1,626
Price-SE 1,309 1,341 1,521 1,517
Price-We 1,119 1,260 1,378 1,578
Table 2: 2006 Demand Volumes (metric tons) & Prices ($/ton) Over 4 Quarters
The challenge for the SMLT was clear enough: Should they engage in risk hedging activity for their plastics resins, possibly including taking positions in crude oil? The analysis presented by the Plastics Team (see Exhibit 2) seemed to suggest that even a small portion of about 1% of total spend on HDPE could lead to significant reductions in maximum expenditures on HDPE in North America (and, by extension, to other plastics resins and other regional markets).
But there were a number of central issues yet to consider. Was this financial approach to hedging cash flows the right approach for the Unilever Supply Management group? Should the same approach be extended to other resins and, if so, with what benefits? Should the SMLT be the one to implement this approach or should it be done by Unilever Treasury? What controls should be put in place to make sure that the hedging that was done was limited to the specific purpose of improving SMs performance, and not for speculative purposes? What benefits, if any, would information provided by this hedging strategy bring to improving buying performance and contracting with respect to ULs plastics purchases?
As Uwe Schulte went through the materials for the coming SMLT meeting, he couldnt help but think this was a whole new game as far as supply management was concerned.
609-006-1
Copyright 2009 INSEAD 8 01/2009-5563
Appendix 1
Risk Mitigation Instruments and Supply Portfolios2
Call and Put Options
A call option gives the holder the right, but not the obligation, to buy the spot asset on or
a before a predetermined date (the maturity date) at a certain price (the strike price),
which is agreed today. Call options come in various flavors. For instance, in American
options, execution is allowed from a given execution/start date through any time up to
and including the expiration date. By contrast, a European option can only be exercised
on the maturity date itself. Figure 8 illustrates the potential payoff from a European call
option.
A put option is similarly defined as giving the holder the right, but not the obligation, to
sell the spot asset on or before the maturity date at the strike price.
Figure 8: Payoff of a European Option
SWAPS
A swap is an agreement whereby a floating (market or spot) price is exchanged for a fixed price over a specified period. A swap buyer pays the fixed leg and receives the floating leg. A swap seller pays the floating leg and receives the fixed leg. Swaps are financial agreements but they essentially assure (for the contracted volume of the swap) that the swap buyer will pay the exact price of the swap for the commodity in question. The effectiveness of swaps is summarized in Figure 9.
2 For further details on available instruments and risk management strategies, see Aswath Damodaran,
Strategic Risk Taking, Wharton School Publishing, 2008.
-
-
0
2
4
6
8
10
12
0 2 4 6 8 10 12 14 16 18 20Terminal Price
Pay
Pro
Stri
Price of the Option
Payoff/Profit
Probability of
Terminal Price
609-006-1
Copyright 2009 INSEAD 9 01/2009-5563
Figure 9: Using Financial Swaps for Risk Hedging
The Supply Portfolio Problem
The central question to be addressed in supply portfolio management (for example for HDPE for ULs North American market) is what mix of supply contracts, options contracts and swaps will provide assured physical supply of needed inputs for its production, together with appropriate financial hedges for the associated cash expenditures for these inputs. Physical supplies can come either from pre-qualified sellers or directly from various spot markets. In supply management for commodities, different grades and specifications for commodities often require prior contracting and procurement relations with pre-qualified suppliers. These alternative situations give rise to various forms of commodity risk management, as shown in Table 3 below.
UL Spend
for Plastic
Resins
Floating Index
e.g., for
Crude Oil
Crude Oil Swap at Price Ps
This crude oil Swap gives rise to basis risk as an imperfect hedge against plastic resin price fluctuations
This Swap is a complete hedge against crude oil price fluctuations
Suppose
there is
correlation
609-006-1
Copyright 2009 INSEAD 10 01/2009-5563
Description of Context Instruments used in Optimal Portfolio
Examples
Cost and access differences small and only standard commodities are sourced
Bilateral contracting and financial hedge instruments are defined on a common market and optimized jointly
Energy
Commodity metals
Cost and access differences are large and only standard commodities are sourced
Bilateral contracting used for most physical procurement, with spot market used for topping up supply, and for financial hedge instruments
Logistics services (standard air and maritime cargo)
Fed-cattle (beef), hogs and lamb markets
Non-standard commodities are sourced, but their prices are highly correlated with those of standard commodities
Bilateral contracting used for all physical procurement, with financial hedge instruments, defined on correlated standard products, used as an overlay for hedging
Plastic resins and commodity chemicals
Table 3: Alternative Contexts for Commodity Risk Management of Supply
The standard problem of commodity sourcing and hedging for a large buyer like UL can be stated as maximizing expected profits, subject to physical delivery constraints and some risk constraints. The general structure of the Supply Portfolio Problem (SPP) is as follows:
Supply Portfolio Problem (SPP)
Maximize Expected Profits (where the decision choices are the amounts to contract for from each available physical and financial contract)
Subject to:
Physical delivery constraints (to assure delivery of needed inputs)
Financial risk constraints (on maximum exposures or on allowable losses from financial instruments used for hedging)
Constraints defining the instruments themselves (puts, calls, swaps, contract parameters such as minimum take provisions and flexibility bands, etc.)
609-006-1
Copyright 2009 INSEAD 11 01/2009-5563
This optimization problem is not solved once and for all but on a continuing basis. At the same time, demand uncertainties are resolved, as well as spot prices and contract prices. To the extent that contracts allow flexibility in execution (e.g., call or put options), these are executed to optimize profits on the day by executing all options that are in the money or needed for physical fulfillment. This problem on the day can sometimes be interesting, but in theory it is straightforward and solved by some computer-based algorithm that picks the best options on the day to execute for both physical coverage and financial return. The more interesting problem, which requires both judgment and computer support, is the medium to long-term, on-going Supply Portfolio Problem (SPP). Various forms of the SPP have been developed for various types of markets, and the details of these differ considerably across these markets. Except in very simple cases, the solution to the SPP must be accomplished using Monte Carlo simulation (together with a simulation optimization engine). We illustrate this below for the plastics problem for UL North America.
609-006-1
Copyright 2009 INSEAD 12 01/2009-5563
Appendix 2
The Initial Risk Management Plan for Plastics
Jan-Jelle Van der Meer, responsible for procurement of food ingredients for the SMLT, had taken a strong interest in the commodities risk management project and Uwe Schulte had charged him to guide the initial effort on plastics. Van der Meer had a doctorate in chemical engineering and over several months of preparatory work had overseen the development of an analytic framework to guide the Plastics Team. The resulting framework entailed the following analysis process:
Start with product and market structure and determine the pattern of procurement and correlated markets that could provide risk management hedges.
Analyze historical data to obtain relevant random variables (demand, price, and correlations) and the associated predicted Spend (total expenses for sourcing and delivering the relevant quantities to production sites).
For a given pattern of procurement choices, which gives rise to the unhedged probability distribution of Spend, analyze risk management overlays that could (at a cost) reduce right tail spend-at-risk (SaR) exposure.
Determine the efficient frontier that trades off increased total Spend against decreased right hand tail exposure or SaR.
As an example of this process, the Plastics Team presented their results for ULs North American expenditures on HDPE. The team used 2006 market data throughout, but they assumed they were at the beginning of 2006 and planning a procurement strategy for that year. Alternative portfolios were evaluated in terms of the total expected spend on HDPE, including the cost of any hedge instruments used. Also of interest were exceedance probabilities for various upper limits (or targets) on total annual spend for HDPE-NA.
A simple simulation in Crystal Ball was constructed to evaluate various risk hedging strategies for the NA HDPE spend. Below are the results for the fourth quarter of 2006, based on the mean values of price and demand quantities at the beginning of 2006. Table 4 shows the assumptions underlying the simulation (all distributions were tested and found to be well approximated by the log-normal distribution). Table 5 shows the results of using just HDPE call options and Table 6 shows the results of using HDPE call options plus Brent crude oil swaps in the indicated amounts. In each case, both the hedged and unhedged (expected value of) Spend are shown. The unhedged value is simply in the cash outlay for HDPE by UL in the market. The Hedged Spend is this cash outlay for procurement adjusted by the cashflows (positive or negative) resulting from the hedge instruments purchased. A number of other combinations of calls and swaps could also be considered, but these two examples illustrate the general consequences of hedging in this case, which can be summarized as follows.
Hedging costs money: The expected value of Spend will be greater than if one did not buy hedging instruments because (on average) no financial broker or investor will take the other side of these market instruments without some expectation of profit. Note, for example, in Table 5 below, that the expected cost of the HDPE call options is $360,000. Of course, the large standard deviation of the value of these call options also tells us that they are in the
609-006-1
Copyright 2009 INSEAD 13 01/2009-5563
money some of the time (and given their structure, we know that this occurs when the price of HDPE is high and therefore these options help to offset high HDPE prices).
Hedging cuts off the right hand tail of the spend distribution: Both the standard deviation and the probability that (Spend + Hedging Costs) exceeds a given target threshold will decrease as hedge instruments are purchased. This is, of course, the primary purpose of hedging. Note, for example, in Table 5, that the HDPE call options reduce the probability of exceeding the target spend figure of $100 million from 0.093 to 0.076. They also reduce the probability of exceeding the target spend figure of $110 million from .013 to .006. There are further reductions in the furthest extremes of the spend distribution since it is precisely for these values that the HDPE call options are clearly in the money. (To see these reductions in the extremes, check the distribution of Spend versus Hedged Spend in the Crystal Ball output.)
The effectiveness of swaps depends on their correlation with the underlying spend: For example, the assumption of 0.65 correlation (not that high!) between crude oil and NA HDPE prices means that crude oil swaps are not that good a hedge for HDPE price volatility in the present case. Indeed it is clear that the crude oil swap hedge here is actually less effective than the HDPE call options alone in reducing the right hand tail of the HDPE Spend distribution (compare 0.76 vs. 0.78 in Tables 4-5 in reducing the exceedance probability for a Target of $100 MM), primarily because these swap options are too expensive relative to their risk hedging benefits (note that buying 50,000 swaps at the indicated swap price has an expected cost of $220,000). Of course, if a more attractive swap price were available, then such swaps could play a role in an efficient hedging strategyhere they clearly do not.
Mean Std Deviation
UL Demand for NA HDPE Qtr 4 55,000.00 4,400.00
Crude Oil Price/Barrel Oct 1 $68.00 $6.00
HDPE Price ($/ton) Oct 1 $1,594.00 $106.00
Correlation of Crude with HDPE 0.65
HDPE Option Price/Ton Oct 1 Calls $28.25
HDPE Execution Price/Ton Oct 1 Calls $1,650.00
Crude Oil Futures/Swaps ($/Barrel) $72.50
Table 4: Assumptions on 4th Quarter HDPE Spend Calculation for UL NA
609-006-1
Copyright 2009 INSEAD 14 01/2009-5563
Mean Std Dev
Unhedged Quarterly Spend ($ Million) 87.74 9.23
Hedged Quarterly Spend ($ Million) 88.10 8.24
Value of Crude Oil Swaps ($ Million) 0.00 0.00
Value of HDPE Call Options ($ Million) -0.36 2.25
Prob{Unhedged Spend > $100 Million} 0.093
Prob{Hedged Spend > $100 Million} 0.076
Prob{Unhedged Spend > $110 Million} 0.013
Prob{Hedged Spend > $110 Million} 0.006
Table 5: UL HDPE NA 4th Quarter 2006 Spend (as predicted on 1/1/2006) Hedging
Strategy: 50,000 Call Options in HDPE CMAI NA Spot Underlying
(Prices per Table 4)
Mean Std Dev
Unhedged Quarterly Spend ($ Million) 87.74 9.23
Hedged Quarterly Spend ($ Million) 88.32 8.25
Value of Crude Oil Swaps ($ Million) -0.22 0.30
Value of HDPE Call Options ($ Million) -0.36 2.25
Prob{Unhedged Spend > $100 Million} 0.093
Prob{Hedged Spend > $100 Million} 0.078
Prob{Unhedged Spend > $110 Million} 0.013
Prob{Hedged Spend > $110 Million} 0.006
Table 6: UL HDPE NA 4th Quarter 2006 Spend (as predicted on 1/1/2006) Hedging
Strategy: 50,000 Call Options in HDPE CMAI NA Spot Underlying and 50,000 Crude
Oil Swaps on NYFE (Prices per Table 4)
609-006-1
To order INSEAD case studies please contact one of the three distributors below:
ecch, UK and USA Centrale de Cas et de Mdias Pdagogiques
ecch UK Registered Office: ecch USA Registered Office: www.ecch.com www.ecch.com www.ccmp-publishing.com Tel: +44 (0)1234 750903 Tel: +1 781 239 5884 Tel: 33 (0) 1.49.23.57.25 Fax: +44 (0)1234 751125 Fax: +1 781 239 5885 Fax: 33 (0) 1.49.23.57.41 E-mail: [email protected] E-mail: [email protected] E-mail: [email protected]
Boulevard de Constance, 77305 Fontainebleau Cedex, France. 1 Ayer Rajah Avenue, Singapore 138676 Tel: 33 (0)1 60 72 40 00 Fax: 33 (0)1 60 74 55 00/01 www.insead.edu Tel: 65 6799 5388 Fax: 65 6799 5399 Printed by INSEAD
609-006-1