GM and China: Exemplifying the Need for Event Driven Research

Something I learned from a conference on China today, featuring  David Hale of Hale Advisors LLC and ChinaOnline:

There are currently about 140 million autos operating in the United States today.  That figure is expected to be about 280 million by 2050. 

There are currently about 12 million autos operating in China today.  That figure is expected to be about 500 million by 2050.

Having cited the above information from a Goldman Sachs auto analyst's report published earlier this year, Hale went on to comment, "It is difficult to imagine where the world will find either the base metals or gasoline to accommodate such a dramatic expansion of he car population in China and India. Indeed, it would be an exaggeration to suggest that China and India will force the world to switch to widespread use of fuel cells at some point during the next twenty years."

Meanwhile, Bob Lutz, a GM Vice Chairman, just posted a "contrarian viewpoint" with regards to its stance on introducing its next generation of full-size trucks:

"I’ll admit that on the surface it may seem incongruous to introduce vehicles like this, given today’s fuel prices. But, I have to tell you, these products still make a lot of sense."

GM began developing these trucks three years ago with oil below $20/barrel.  Is it contrarian to roll-out vehicles based on three-year old economics, or is it more like they have no alternative?

Perhaps its the ladder given that a BusinessWeek article published last year mentioned that GM was already planning on selling hybrid SUVs in 2007 and had teamed up with DaimlerChrylser to defray costs of developing the technology.   

And coming to the point highlighed by the title of this post, GM is either a value play or an unwise investment.  The trick is to figure out which one, and that is where some decision tree analysis would come in -- which has nothing to do with the detailed financial forecasting models I've posted elsewhere on this blog.   

So here is one speculative branch of the decision tree:  Perhaps an investment from China a la Lenovo or (the attempted) Unocal is the next logical step to look for, which could make it a win for value investors.   

A good buy side auto analyst should have already assigned a probability to this and related events occuring, as well as its the possible financial impact on the company and its stock given different potential levels of investment.  (And this is where tools like MindManager combined with Excel come in -- I intend to post an example of this, hopefully soon).

Unfortunately this kind of event-driven research is not often produced by Wall Street -- for it takes imagination as well as disciplined number crunching (and time).  But if I'm wrong, please point me in the right direction.  I'd like to read and, ultimately, produce this kind of research myself. 

Comments

[martha]

Interesting post. The problem that I have is that anyone who looks out more than five years is crazy. Projecting auto demand in a linear fashion is beyond childish. Seriously, with this post you seem to be breaking from many of your excellent observations - forecasting over long periods of time and assuming linearity.

[Steve Castellano]

You are right, in a way. In the 1950s nuclear power was the holy grail to the American economy, and 50-year forecasts showed a nuclear power plant in every city, town and villa. So much for the accuracy of those forecasts.

But that was forecasting the emergence of a technology. Forecasting bandwidth-driven revenue among the telecom services companies more recently was also a forecast of transitioning technologies. Forecasting revenue based on transitioning technologies is difficult at best, a point made obvious to anyone alive post March 2000.

But forecasting GDP growth of a country is a lot easier -- that is, there are less pitfalls that will interfere with your forecast. So lets look back at 50-year forecasts of the American economy in the 1950s. Whatever assumptions the forecasts were based on, they probably showed some kind of increasing growth (linear, stepped, accelerating -- irrelevant. It was up).

Auto production is a function of GDP growth and more specifically perhaps per capita consumption. We can look at China and see that its demographics and current per capita spending levels suppport the idea that this country has a long ways to grow.

Assuming future expected economic growth, perhaps make an ultimate worse-case estimate. For example, assume that 90% of the Goldman forecast, whatever its asusmptions (I have to assume they are logical since they come from GS) is bogus, that leaves a 2050 figure of 50m estimated autos in China.

In of itself, a worse case estimate illustrates the increasing demand for associated raw materials around the world, driven by China, India and other developing nations. Whether its 50m or 500m by 2050 is not as important as the point it is illustrating -- that future growth is up, and significant.

On the other hand, there are no doubt experts on China, India and autos that are probably make the case that these emerging powers will implode from any number of economic or social crisis. So lets go back to the 1950s 50-year forecasts in the U.S. Perhaps those forecasts should have been dismissed because there was risk of communist uprisings or nucelar war. That there were risks doesn't take away from the need to make an assessment on the future.

So you are right, forecasting more than 2-3 years with any level of detail is suspect. But, to also arbitrarily limit any kind of logical, extraoplated visibility down the road is just as short sighted.

In short, forecasting a function of country GDP is much easier, and more of a high-level exercise, than forecasting the emerging dominance of any particular technology as the key driver of a company's revenue-- though there are dangers to each.