Streetwise Professor

December 1, 2011

Just In Time for Christmas!

Filed under: Commodities,Derivatives,Economics — The Professor @ 6:12 pm

My new book, Commodity Price Dynamics: A Structural Approach (Cambridge UP) is now available.  Only 3 left in stock at Amazon!  Order Soon!  Sorry–no Kindle version.

From the book description:

Commodities have become an important component of many investors’ portfolios and the focus of much political controversy over the past decade. This book utilizes structural models to provide a better understanding of how commodities’ prices behave and what drives them. It exploits differences across commodities and examines a variety of predictions of the models to identify where they work and where they fail. The findings of the analysis are useful to scholars, traders, and policy makers who want to better understand often puzzling – and extreme – movements in the prices of commodities from aluminum to oil to soybeans to zinc.

I’d quote from the blurbs, but I’m just too darned modest.

In all seriousness, the book represents the culmination of years of my professional research.  I’m currently working on some extensions that apply the techniques utilized in the book to some current policy issues, such as the effects of speculation and commodity pricing anomalies, such as the huge basis observed for wheat in particular (and corn and soybeans too) in the mid-to-late ’00s.

I think the structural approach is quite useful in understanding how commodity prices behave, and crucially, how commodity derivatives should be priced. That’s where the book differs from earlier explorations of the structural approach, like the book of Williams and Wright from the early 1990s.

Moreover, I believe that the structural approach is useful in demonstrating that price behaviors that are commonly considered to be irrational, and driven by excessive speculation or noise trading, are in fact perfectly consistent with rational expectations-based models.

Also, I try to resist the quite common tendency to fall in love with the models one works on. One of the main contributions is to show where the structural model for storable commodities fails miserably, and to suggest extensions that may correct these failures. As I say in the introduction, I try to break the models and learn from the pieces.

For my next book, I’m changing gears, and working on a subject that has been covered in myriad SWP posts. It will be titled “Market Macrostructure” and will also be a Cambridge imprint. The first half of the book will analyze the organization of the execution of financial transactions, focusing on issues like liquidity network effects and dark pools. The second half-will examine the organization of post-trade services, most notably clearing. I’m sure that you are all shocked by that.

Print Friendly, PDF & Email


  1. A little pricey for only 230 pages.

    Comment by John — December 1, 2011 @ 6:33 pm

  2. I had already bought one copy, but the scarcity of the book led me to buy another. The first copy will stay with the Commodities team of BM&FBovespa.
    Congratulations for your book !

    Comment by Marcos Carreira — December 1, 2011 @ 6:35 pm

  3. Thanks for the heads up-now you only have two left.

    I hesitated on something a couple days ago and so I lost what I could have gained. 🙂 I hope your book will correct that type of error. Congratulations on your book.

    Comment by pahoben — December 1, 2011 @ 6:37 pm

  4. thanks for the heads up. I just bought amazon’s last one.

    Comment by vbounded — December 1, 2011 @ 7:03 pm

  5. Amazon is offering a $4.74 gift card for your old copy. I’ll top that with a cool $5.00. Any takers?

    Comment by ThomasL — December 1, 2011 @ 7:52 pm

  6. Thanks, folks. Very gratifying.

    But John–it’s 230 pages bursting with insights! And PDEs! Seriously: welcome to the world of academic publishing.

    The ProfessorComment by The Professor — December 1, 2011 @ 9:24 pm

  7. I think the price if pretty decent and I am seriously considering buying Do you do any comparisions between the structural factors you obtains with those of Schwartz et al?

    Comment by Surya — December 2, 2011 @ 7:47 am

  8. Good question, Surya. Yes. Yes I do. A major theme in the book is using structural models to identify the deficiencies in reduced form models a la Eduardo’s.

    The ProfessorComment by The Professor — December 2, 2011 @ 9:47 am

  9. More specifically, I show in particular how the reduced form models will do a bad job in pricing swaptions or calendar spread options b/c they don’t get the correlation structure right.

    The ProfessorComment by The Professor — December 2, 2011 @ 9:48 am

  10. Congratulations Craig.

    Comment by Jim — December 2, 2011 @ 9:49 am

  11. congratulations – Options pricing, particularly on SWAPS and any term structure related instruments has been a major issue of late – While most are familiar with using the OIS rates for pricing and marking of LIBOR swaps – they were still using implied volatility from the regular market in options pricing! Any intrusion of real world considerations is deeply needed.

    I would love to hear your comments on OAS models for option impacted debt. They have always seemed somewhat tautological to me, and the method of developing an option price as yield, particularly as the yield itself is impacted by optionality, rather than a dollar price reflecting optionality is counter-factual, to say the least (except in some CDS, one usually pays an upfront premium!).

    Comment by Sotos — December 2, 2011 @ 4:14 pm

  12. Well it’s on my list, but I’m probably not gonna buy it until it gets a bit cheaper.

    Comment by John — December 2, 2011 @ 9:55 pm

  13. Congratulations! I pre-ordered my copy. Note: Only $63 on, what gives?

    Professor: Will there be a student budget-friendly softcover copy in the future?

    Comment by Jack — December 3, 2011 @ 11:56 am

  14. Thanks, Jack. Price discrimination! I guess demand is more elastic in Canada. But given the ability of US folks to shop via Amazon, you wouldn’t think that was a viable pricing strategy. Interesting.

    Softcover availability will probably depend on how well the hardcover sells. I’ll ask my editor about it.

    The ProfessorComment by The Professor — December 3, 2011 @ 3:52 pm

  15. Ah. But I see you’re Canadian, Jack 🙂

    You mention “student-budget friendly”–unfortunately, that hardly seems to be a major consideration for publishers. Textbooks are outrageous.

    The ProfessorComment by The Professor — December 3, 2011 @ 3:55 pm

  16. Thanks Prof. Getting the correlation structure right is pretty interesting. I thought the correlation structure decays with increase in time separation between forward contracts. This would give rise to a e^-|i-j| type correlation structure most of the time. Is that too simplistic?

    Comment by Surya — December 3, 2011 @ 9:30 pm

  17. Yes. Too simplistic. Depends on the shape of the forward curve. Correlations approx 1 when market is in full carry. Correlations can be quite low when market is in extreme backwardation. Storage is what connects prices along the curve. When storage is high=>contango=>high correlations. Storage low=>backwardation/low correlations.

    Very hard to get a reduced form model to do that trick.

    The ProfessorComment by The Professor — December 3, 2011 @ 10:51 pm

  18. I know reduced form models are all the more popular in interest rate option markets and factor LMMs are often PCA based. Is it possible to construct structural factors for these along the lines of what you have done for commodities? Might be very useful …..

    Comment by Surya — December 4, 2011 @ 12:29 pm

  19. @Surya–One of the things that attracted me to commodities is that the fundamentals are relatively transparent, and can be modeled relatively realistically in a supply-demand framework. I am skeptical that would be possible for IR models. Cox-Ingersoll-Ross is an equilibrium model for a very simple economy. Not saying it’s impossible, but given computational constraints, I think any computable model would be far too abstract, and any relatively concrete model non-computable.

    The ProfessorComment by The Professor — December 4, 2011 @ 2:05 pm

  20. Just to add the cherry to your last post, professor, IR forecasting models have been notoriously inaccurate. IR models focusing on evaluations, particularly those relating to spread and risk adjustments over time, have tended to the stochastic/Monte Carlo types- lately using a Markov process for predicting volatility shifts. This seems like a mathematical expression of the old joke about predicting often. Unfortunately these have proven themselves very iffy in practice (if anyone says no data for parametrization again I will scream) . Another way of saying that the fundamentals of IR is unclear is that a lot of the “markets” tend to disappear and the pricing data can be truly horrendous. I am sure it is very hard to model supply and demand when even accurate pricing for only the most generic instruments is remotely accurate.

    A brief story. In the early 1990’s I was hired to build a commercial pricing engine to price Agency ARM MBS – no hte most generic but still common securities – for back office marks. This was done using “trader Heuristics”: Basically modelling how a trader would look at a pool, scratch his or her privates, grunt and say Plus 5 (ticks) or minus 10 bps at PSA 650, etc. OAS didn’t work, and often comparative yields didn’t work due to pool specific behavior. In other words this was not simple but hardly rigorous – a lot of the pools had been securitized for Basel I purposes (20% asset weighting versus 50% for SF mortgages)or for collateral, and had some really weird stuff in them. But again the service was for the Bank back offices, so if we got reasonably close on the majority of stuff, it was OK.

    Having launched this thing in 90 days, I was stuck running it. A few days after launch I get a call that so and so BIG NAME Mutual fund wants to discuss our pricing. Spitting out my coffee, I ran upstairs to my friend the marketing and sales manager (who had gotten me into this mess) and asked him loudly what the f*ck he thought he was doing – were we going to use this crap for NAV’s? We’ll all end up in court!

    His reply was “Sotos, your prices may be terrible, but Merrill’s are much worse.” He was right then, and still is today.

    Enough bloviating.

    Comment by Sotos — December 4, 2011 @ 4:57 pm

  21. Funny but tragic at the same time, Sotos. And think how much worse it is today, with the proliferation of instruments nd complexity.

    The ProfessorComment by The Professor — December 4, 2011 @ 5:00 pm

  22. But the one eyed man is king in the land of the blind.

    The ProfessorComment by The Professor — December 4, 2011 @ 5:01 pm

RSS feed for comments on this post. TrackBack URI

Leave a comment

Powered by WordPress