By on April 18, 2013

Despite a chorus of largely uncritical reporting, there is a growing contingent of those who question VW’s claim that their new MQB modular architecture will bring about significant savings. The latest among them is Bernstein Research. A report by noted analyst Max Warburton (who recently authored the definitive study on Chinese cars) questions VW’s claims and shows how MQB may be helpful, but not nearly as significant as VW claims. On the other hand, it’s worth noting that these reports are meant for end-user investors, not necessarily industry types.

In theory, the MQB platform is a dream come true for any auto manufacturer, allowing a one-size-fits-all platform to span across their expansive product offerings and further increase economies of scale by offering both standardization and flexibility all in one neat package. From that perspective, it’s easy to wonder “how can Volkswagen NOT save money?” And for a company like VW, which is arguably an industry leader in scale, 20 percent might even be an appropriate  figure.

Like any prudent business student, I began to question the numbers being disseminated from the company. 20 percent seemed too neat, too easy, and most of all, too tailor made for a largely unquestioning automotive press. My thoughts trailed back to my introductory finance class. One key consideration that should be taken into mind when analyzing company valuations is the optimism of those responsible for creating the financial models. Naturally, I figured 20% was the result of this phenomenon. If you think about it, it’s tough to imagine, “VW Announces MQB to Save 2% of Costs,” making a splash in the headlines.

Bernstein Research’s analyst report on MQB presents a more contrarian view. Overall, Bernstein was unimpressed with MQB, claiming that the benefits were “over-hyped.” This viewpoint had differed so drastically from my initial research on modular platforms, which hailed them as the next wave of change and innovation in the automotive industry. How could this gap be reconciled?

To fully understand and appreciate the report, I had to get into the mind of its intended end-user, an investor. The mindset of an investor is much different than that of a car enthusiast. While industry observers herald MQB as the next great leap forward for automotive manufacturing, an investor cares about one thing and one thing only: money. From an investor standpoint, if MQB cannot deliver any value though its stated cost savings and  result increase company earnings, it is largely irrelevant.

The report states, “There may be some savings from this [MQB], but they are likely to be very modest.”[1]  The modest nature of these savings stems from a number of sources. First, MQB does not help to lower a manufacturer’s major costs like raw materials, labour, and assembly hours. Remember that economies of scale, at its core, is the ability for a business to save money by spreading its fixed costs across a higher volume of production. This definition would lead to the assumption that by standardizing all of its platforms, VW will be able to cut costs throughout the supply chain by having their suppliers reach economies of scale. However, the report argues that past a production level of 1 million units, there begins to be diminishing returns.

Bernstein explains, “scale economies are usually exhausted at the plant level rather than the firm level.”[2] Since most OEMs are already at full scale, they already purchase parts from suppliers in some sort of economic batch size. In order for suppliers to save through economies of scale, they would have to increase their production. Since suppliers are also already at full scale, they would have to increase the capacity of their plants in order to increase their output. The capital expenditure related to increasing capacity, whether it is machines or labor, effectively negates any sort of cost saving that was originally sought after.

Theoretically, if VW sourced its parts and raw materials from one centrally located monster plant, it would be much easier to realize these savings. VW’s geographical diversification and localized sourcing makes this much harder to achieve, as each separate location has its own associated fixed costs. Think about it in terms of trying to house all of your cars in one mega garage, or in 10 separate garages across the city. The one mega garage has a single set of maintenance and utilities, which the latter option has 10.

A more grounded figure for MQB’s cost savings is somewhere in the neighborhood of 2%-5%, similar to what Renault-Nissan claimed. As mentioned above, 2% is not the most attractive number to unveil to the public, but this relatively small figure should not be dismissed completely. VW’s costs of sales were over 157 billion euro in 2012. A 2% cost saving would translate to an additional 389 million euro in earnings before interest and taxes, or an addition of 0.83 euro on a per share basis. Even on the lower end of estimations, it is clear that if MQB can help lower costs by any amount, it still helps to deliver value to VW and its shareholders. If you were to put $389 million in my pocket, I wouldn’t be too upset.

When considering the competitive nature of the automotive industry and the resulting intense margin compression that manufacturers face, the amount of cost savings are even more impressive. I like to think about it in terms of a Formula 1 race, in which the shaving of even a few milliseconds off of a car’s lap time is a huge accomplishment. Success is relative, and numbers, whether they be tenths of a second or percentage points, are meaningless without context.

Even if savings derived from MQB are under 2%, let alone 20%, it still provides a huge strategic advantage for a multinational auto manufacturer like VW. The ability to remain flexible across its product offering will allow VW to continue to serve existing and emerging markets, each with its own specific needs. It will be interesting to see if VW can effectively leverage MQB’s capabilities in its implementation. The ability to “flex” plants quickly (from a Golf to a Tiguan for example) based on market demand is one aspect that has been barely mentioned, but could prove to be a significant advantage.

Graeme Kreindler is an HBA Candidate at the Richard Ivey School of Business at The University of Western Ontario. 

TTAC thanks Max Warburton and Bernstein Research for providing the report.



[1] Pg 30

[2] Pg 24

Get the latest TTAC e-Newsletter!

19 Comments on “EXCLUSIVE: Bernstein’s MQB Report In Full...”

  • avatar

    While I agree that the 20% cost savings is suspect, the architecture advantages are greater than just simple parts savings:

    1. Shorter design / testing cycle
    2. Significantly faster assembly time
    3. Ability to run multiple MQB models off the same line
    4. Less training due to more commonality
    5. More flexibility in line staffing, production, etc.
    6. More common tooling

    Perhaps VW is calculating 20% taking reduced development, assembly and tooling costs into consideration?

  • avatar

    Excuse me but “raw materials, labour, and assembly” are not fixed costs. Make fewer cars and you need less of those items, not the same amount. Development, tooling and the plant are fixed costs.

    • 0 avatar

      I think what was meant was that the costs are the same no matter if you are building a modular platform car, or building a dedicated design.

      (If true, that is sloppy use of wording however.)

  • avatar

    I think there is a significant misunderstanding about this 20% figure: it is not 20% of the total cost, but 20% of the *manufacturing* costs, and a 30% reduction in manufacturing time. That’s not the same as saving 20% of the *total* cost of the vehicle. And I believe Bernstein is caught up in this same misunderstanding.

    Some more interesting MQB reading here:

    And on MQB and lightweight construction:

    Automotive news review of MQB:

    Winterkorn speech talking about MQB savings:

    • 0 avatar

      I was thinking the exact same thing based on previous articles I’d read. I was under the impression it was 20% reduction in production cost rather than total cost.

      • 0 avatar

        Using a common “modular” platform does little to reduce the production cost it’s biggest benefit is spreading the development costs across more sales. In theory yes you could produce more models on the same assembly line and need fewer assembly lines but you are limited in the amount of vehicles you can make on a single assembly line so to really get the bang for the engineering funds you’ll need more assembly lines and of course buyers.

        • 0 avatar

          If MQB makes assembly simpler — and here Winterkorn says it can reduce the assembly time by up to 30% — then it will reduce the manufacturing (assembly) costs as well. The same factory, with the same set of employees, will be able to build more cars in the same amount of time. And that means a lower cost per vehicle.

          Of course you need to have buyers for all the cars you build. But that’s true whether you’re using the MQB architecture, a conventional platform, or hand-building cars with ash.

          • 0 avatar

            I’m not sure how a modular platform can reduce assembly costs that much. It’s all been reduced to steps that all take very close to the same set amount of time at a given station.

            So I don’t see how the commonality will make it so that it takes less time to do the prescribed task at EVERY station along the line.

    • 0 avatar

      That was my understanding too.

      The total cost of a vehicle is much more complicated — even using a similar platform still requires some degree of engineering, and there will be some parts that aren’t common.

      Direct quote from one of the links above:
      “Improving the automobile plants towards MQB system will certainly incur a large amount of price but concurrently it’ll reduce the price of manufacturing by almost 20% and reduce the assemblage times by a remarkably 30%.” [sic]

      • 0 avatar

        Part of me really wonders about the ability of the MBQ production to really deliver on all the promises. It seems like the approach that’s unique here is that they’re using a modular platform as a basis for improving the flexibility and efficiency of manufacturing, not just development. In other words- before you didn’t have to develop a new platform but your G8 was still produced on a different line than your camaro. Now VW can make a passat OR an A4 on the same line depending on demand. Actually- from what I understand from the Siemens guys- VW’s already been working towards this. I think the change here is that now instead of all the C5 cars all made on the same line, ALL the cars CAN be made on the same line.

        However, the more different products you make one one line the more complex every step of the process becomes. It’s not just about having more sophisticated controls or production staff. Your ERP systems have to become more complex to deal with the more complex planning required or you carry to much raw materials. Your quality has to become more sophisticated to keep up. Changes made to the line for one product can have a substantial impact on the rest of the line. Qualification of new equipment becomes a nightmare. It’s like trying to use excel to solve some problem that uses linear equations. For a 2 or 3 order system it’s pretty helpful. If you try to use it for a 6 or 7 DOF equations there’s just way to much going on to even keep track of.

        After giving it some thought I wouldn’t invest in VAG right now. It may look promising- if anyone can pull off platform sharing at this level it’s VW. But as a 2.7t c5 owner- I can tell you that sometimes the germans have a way of making things a bit too complicated for their own good.

        • 0 avatar

          Wasn’t one of the pitfalls of the 2.7t, as used in the S4 and A6, that the germans didn’t build it complicated enough choosing to omit adequate cooling for the heads/valve train.

          • 0 avatar

            No. The cooling works fine. The issues start with the radiator. Which takes 5 hours to access (have to take the front of the car off- ask my why the bumper lines don’t meet up quite right with my headlights).

            Another great example is the front suspension: there are 8 control arms… on the front. And struts. There are so many bushings I can’t even keep track. The roll bars have a sensor (which is prone to breaking during axle swaps… of ripped CV joints). The axles bold to the trans with 12 point bolts. Sockets for which don’t actually exist. Even the wheel hub end of the axles bolts to the suspension uprights with a 19mm allen bolt. Yes- I now have a 12 point set and a 19mm allen socket. I know- you’re thinking “12 point- what a moron it’s just a torx!” You’re wrong.

            Oh yea- they switched the control arm design half way through the model year- some have aluminum control arms, some are steel- they look the same. The only difference is the radius of the ball joint end. Again- ask me how I know.

            The alternator on my 93 jeep is on the top of the engine. 1 hour to change as a 20 year old who just bought wrenches. The alternator on the audi is underneath… THE WHOLE DAMN CAR!!! You have to unbolt the front “Lock carrier” (front bumper, radiator, trans cooler, oil cooler, etc…) to change… THE ALTERNATOR!!!!!!!

            Also… The timing belt…

            It just makes me think of what it must be like to do maintenance at these plants… “O well jetta production has been shut down because we need to change a bearing in the press for the door sills. Well no we can’t use a standard bearing because the phaeton requires tight tolerances so we have to have it custom made from switzerland and they’re all on vacation for a month because it’s Europe… Yes we had to get the bearings for our equipment in south Carolina made in Europe- Americans can’t make things!”

            What is great about c5 audis is that they’re reliability makes them cheap as hell to buy used for how much car you get. BUT… There’s a reason for that…

            Lastly: I am convinced there is a dude named Hanz who works for audi. He designs a lot of non-critical components- placement of bolts, etc… His hands are teenie tiny. Hanz doesn’t understand why we can’t get the axles off- he had no problems!

    • 0 avatar
      Athos Nobile

      Thanks for those links.

      • 0 avatar

        Interesting- this intersects with my concerns about this sort of common architecture. What happens if (when) you get something wrong? It’s one thing if a major problem emerges on one model line, but what happens if it occurs right across the shared component set? Do all your cars end up with, say, faulty brake pedal assemblies?

  • avatar

    Graeme you can do better! Look at the info th009 provides. Thanks btw for the extra info!

    I’m skeptical at the 20%, that is based as I understand on the manufacturing costs. My skepticism is mostly based on what Graeme points to: it sounds to nice.
    On the other hand I have to dig through the Bernstein report and the extra info from th009.

    According to Graeme’s calculation the manufacturing costs at VW are:
    389 million/2%=19.45 billion
    So manufacturing costs are 19.45/157=1.23% of the total costs of sales. Hmmmm, numbers numbers.

    What about cutting manufacturing costs by producing in a country where the currency is low and just ship your cars to where the markets want them?

    • 0 avatar

      According to the study below, the labour costs per vehicle at GM and Ford (in the US) are about $1750, or about 6% of today’s average transaction prices. Let’s say 7% of the FOB price at the factory. VW’s German costs are higher, Chattanooga and Puebla are lower.

      But if you can save 30% on labour costs (remember, 30% reduction in manufacturing time) that is probably equivalent to roughly 20% of total manufacturing costs. In any case, let’s call it a $500 (30% of $1700) savings.

      Multiply that by VW’s planned 10 million units per year, and your looking at $5B in manufacturing cost savings per year.

      Incidentally, that’s not so far from Graeme’s guesstimate of 2% of EUR 157B, or EUR 3.1B (Graeme’s decimal point is off by one position) — that’s $4B at current exchange rates.

      • 0 avatar

        Thanks again!
        Than it seems that perhaps the magical 20% can be done. I can imagine all sorts of benefits in tooling/refitting factories etc etc. But I don’t see Graeme digging into this.

        Bit of a problem is that everybody focusses on that 20% without actually clearly stating on what it is based, Graeme included. And then what I also do not see is back of the envelop calculations like yours.

        Some of the things he says are simply not true:
        “First, MQB does not help to lower a manufacturer’s major costs like raw materials, labour, and assembly hours.”
        If I can use more engineering time to optimize high strength steel instead of normal steel I can use less materiel. Same for labour and assembly hours. That’s what I conclude from the information of th009.

  • avatar

    If the frame is standardized there is no need to reprogram and/or reposition any robotics involved in assembly.

    • 0 avatar

      But the idea is that they can make different length and width vehicles on the same basic components so the robots will need to have their programing changed. However that is a one time cost and once the program is done for a particular version it is just a matter of choosing that program and running it.

Read all comments

Back to TopLeave a Reply

You must be logged in to post a comment.

Recent Comments

  • mcs: @toolguy: No, it wasn’t his vehicle. In my world, I’ve never seen people buying cars to impress...
  • Arthur Dailey: @KC; I have not personally seen any masks that are just one layer, at the least two layers. if there...
  • Detroit-Iron: F1, like the IOC, supports slave labour, human rights abuses, and validating dictators and other...
  • mcs: One of the things they discovered about the effectiveness of conventional masks with kids is that it kept them...
  • kcflyer: I did find it interesting that the n95 filters particles smaller than the openings in the mask by magicly...

New Car Research

Get a Free Dealer Quote

Who We Are

  • Adam Tonge
  • Bozi Tatarevic
  • Corey Lewis
  • Mark Baruth
  • Ronnie Schreiber