Automotive Value Chain Profit Share
While there is no pre-existing information to fully answer your question, we've used the available data to pull together key findings: Despite an exhaustive search of scholarly, industry, and other trusted media sources, we were unable to find any updated graph to match the original from the 1998 Harvard Business Review article, "Profit pools: a fresh look at strategy." In fact, the original chart appears in presentations as recently as 2015 and the original article is still being cited by current peer-reviewed articles, strongly indicating that an update simply does not exist. Based on the sources that we came across during our research, we believe that it is possible (though far from definite) that one could compile the data from public sources to triangulate a new chart, but that this would be a painstaking and time-consuming process.
Below you'll find an outline of our research methodology to better understand why information you've requested is publicly unavailable, as well as a deep dive into our findings.
We began with a general search for a graph similar to the one shown in the project description, but based on more recent data. Since the original report was published in the Harvard Business Review, we searched Google's Scholar database and found that this article has been cited 238 times by other peer-reviewed works, eight of which were published in 2018. In reviewing the titles and Google's previews of those articles, we found only one which might contain an update to the information, a 2009 article entitled, "Beyond products and services: shifting value generation in the automotive supply chain." However, this article is paywall-locked and based on the available extract, we consider the likelihood that it contains the requested graph to be very low.
Next, we broadened our search to non-scholarly sources. While we located several iterations of the original graph (including many in full color), we were unable to find one attached to more recent data. In fact, we found versions of the original graph being used in presentations dated as late as 2015, strongly indicating that no one had made a more recent attempt to update it.
We next attempted to find one or more public reports which would contain sufficient data to construct a new graph, searching both the abstracts of marketing reports and automotive industry sources such as NADA's 2017 auto dealership report (which may be requested at no cost here; NADA's licensing agreement does not allow us to provide a copy of the report or a direct link to it). We understood that this was outside of the scope of the request but wished to at least determine the likelihood that it could be accomplished in a reasonable period.
While we found bits and pieces of the data (some of which we present below), we did not find any public source with the requested graph or data, nor any combination of two or three sources which would do the same. Indeed, even studying the abstracts of market reports from sources such as IBISWorld, First Research, and Anything Research, we found that even if we had access to the full reports, they would at best contain only parts of the answer. Other, less comprehensive sources, such as this article from Geico which reveals the average profit margin (not the operating margin) for US used car dealerships, were similarly scattered--we could find pieces of the puzzle but could not find even a hint that anyone had pulled together all or most of the information contained in the original graph into a single source.
This is not to say that re-creating such a graph is necessarily impossible. Even though it was outside of the scope of the request, simply searching for an updated graph or similar collection of the data we stumbled upon several of the data points which would be needed if someone were to wish to create an updated graph.
For example, according IBISWorld, the total revenue of all new auto dealerships in 2017 was $905 billion. However, NADA's 2017 report indicates that this amount includes used car sales by those dealerships as well as service & parts. So, for example, the average "new" car dealership in 2017 totaled $34.4 million in new car sales, $18.1 million in used car sales, and $7.1 million in service, parts, and its body shop department. Taken as percentages, this means that 57.7% of revenue was from new car sales, 30.4% was from used car sales, and 11.9% is from parts and service. If we multiply those percentages against the total industry revenue of $905 billion, then we get $522 billion in new car sales, $275 in used car sales, and $108 in parts & service.
For another example of how part of the data could be compiled, according to Statista's most recent report, the US produces 12.2 million motor vehicles per year. The NADA report shows that the average retail selling price for these vehicles was $34,670, with a retail gross profit of $1959. Subtracting the dealer's profit, they were sold by the manufacturer at an average price of $32,711. At that price, the total US auto manufacturer revenue would come to $399 billion ($32,711 * 12.2 million). Since that $399 billion would come from the new car sales part of the value chain, this would reduce the total new car sales to $123 billion, a little under half that of used car sales.
Since creating a new graph is out of scope for this request, we did not pursue this course further--the above are simply the data points that we came across during our research. While we consider it likely (though not definite) that a new graph could be developed from the available public sources, it would be a painstaking and time-consuming process to make sure that all the data is properly accounted for.
Despite a comprehensive search of the available public sources, we were unable to find any which had compiled the data to provide an update to the 1998 Harvard Business Review article. Based on the exact same graph still showing up in presentations dated as late as 2015 and the article being cited in peer-reviewed work as late as 2018, we are confident that there is no more recent work that compiles the same data. Based on some incidental results of our own research, we believe that compiling this data even using only public sources of information may be possible, but would be a difficult and lengthy undertaking.