Jim Manzi does good work for all of us in GM’s Magical Thinking:

I’ve been working my way through GM’s much-heralded restructuring plan that is being submitted to Congress tomorrow. It’s a lot less work than I thought it might be (though I’m blogging late at night as I’m reading it, so all of my comments are subject to revision).
It makes at least one point very effectively: GM has an incredible short-term cash problem. While I am skeptical of bankruptcy in the next 27 days — which GM claims will happen without a $4 billion government cash infusion by the end of the month — it is surely true that at current course and speed, this company is going to run out of money within a few months.
And even worse, if we look out over the next 6 – 9 months, it’s very discouraging, even if they get the bailout. GM has requested $18 billion in loans from the government, in addition to the $8.3 billion they are already getting under the fuel-efficient cars program (which, in practice, will provide a ton of general financial support — I’m willing to bet a lot of money that the green cars program at GM is soon going to find that it simply can’t do without a huge number of accounts payable clerks, market research managers and prime Warren, Michigan office space that just happens to be paid for today out of GM’s general treasury). GM’s base case restructuring plan calls for them to use $4 billion dollars of this $18 billion in December, $4 billion more in January and $2 billion more by the end of March. That’s about a $10 billion drawdown. They also present a far from unthinkable “downside scenario” in which industry auto demand is 10 – 15% lower than in their base case. In the downside scenario, they would draw down about $15 billion instead of $10 billion by the end of March, and have a cash balance at that time of about $13 billion. If industry demand stays at this lower level for the next quarter, they would burn through another $6 billion of cash in the second quarter, and would therefore consume the last $3 billion of the $18 billion they have requested, plus drive their cash balance down by $3 billion, to about $10 billion, or right where it is now. In other words, even if they got everything they asked for and achieved whatever short-term operational and asset sale plans they have, but consumer demand were soft over the next six months, we’d be right back to same place by the end of June.
However, if GM got the loans, and if we have a decent consumer auto purchase market for the next year or two (how’d you handicap that?), and if GM is able to improve its operations sufficiently, then they could squeak by. The point of this document was supposed to be the presentation of the plan to achieve these operational improvements. But there’s no there there. I guess somebody who’s never read a real business plan might mistake this document for one, but it’s a joke. It’s basically a list of assertions of amazing improvements, entirely discontinuous with actual performance to date, that they will achieve. What’s missing is any real indication of how they will go about accomplishing this.
Here are four huge examples:
1. As a practical matter, the core of any GM restructuring plan to achieve sustained cash profitability is unit cost reduction. GM simply asserts that they will do this, and lays out the targets. The only support is a bunch of rhetoric and cherry-picked facts about the recent past designed to lend credibility to these assertions. Consider the chart of historical and projected U.S. active hourly manufacturing costs shown in Figure 3. This total cost has been reduced from $12.8 billion in 2004 to $7.4 billion in 2008. This makes the projected reduction from $7.4 billion to $4.5 billion from 2008 to 2012 seem pretty reasonable — it’s fairly close to a continuation of the pre-existing cost trend. But consider this in context of (as per Table 6) total employment of 167,465 in 2004 and 96,537 in 2008. If we calculate the crude productivity measure of active hourly manufacturing cost / employee, this was $76,434 in 2004. In 2008, it was…$76,655. In other words, the reduction in aggregate active hourly manufacturing costs between 2004 and 2008 appears to have been driven by simply having fewer workers, with unchanged productivity. If we apply this same productivity assumption to the projected headcount of 67,500 for 2012, we get total active manufacturing costs for 2012 of about $5.2 billion, instead of GM’s plan of $4.5 billion. Where does this $700 million dollars of productivity improvement that seems to have heretofore escaped management come from? — No idea.
2. Similarly, look at Figure 4. It is an assertion that so-called structural costs will decline from 34% of revenue today to 25% of revenue within 4 years. That’s close to a nine-point gain in margin in four years, which would be pretty awesome operational performance improvement. Actual structural costs have fluctuated between 30 and 35 percent of revenue over the past five years — so how is GM going to get this huge improvement? They provide one sentence that claims that already-negotiated reductions in legacy costs will do this, and provide a chart in the appendix (Exhibit C-2) that breaks this cost reduction into four big categories. That’s it, as far as I can see. What a coincidence — GM just happens to be at the moment when their fixed cost problems, which have plagued them for decades, are about to go away, if they can just get a bridge loan.
3. In Exhibit B-1, GM notes that last year they forecast that GM would achieve a 2009 unit market share of about 20.6% (3.3M / 16MM), but now project a 2009 unit market share of 22.5% (2.7M / 12MM). Are they really asserting that their competitive position has improved over the past year, in direct contradiction to all external evidence and their own statements elsewhere in the plan? Of course, we have no idea what market share their plan assumes going forward, since they don’t provide this. They also don’t provide any serious financial statements or projections, only a projection for ultimate cash flow by year, without any useful description of the assumptions that generate this forecast.
4. In Table 4, GM claims that total debt will go down from $62 billion to $30 billion, or be reduced by $32 billion. They are asking for a loan of $18 billion. How do they get the extra debt reduction of $16 billion? They don’t say in the plan, but they appear to be basically assuming that their debt holders are going to “restructure” too, which is a 50 cent Wall Street word for agreeing to give up a lot of the debt in the hopes of getting paid at least something. This is like your brother-in-law’s plan to get out of credit card debt by getting Visa to just agree to forgive a lot of what he owes. Felix Salmon made the same observation, and called Troy Clarke, the head of GM’s North America business. Clarke confirmed that GM is assuming exactly this.
Now, the most obvious response to all of this is to say that I’m the fish at this table, because this is not a real business plan, but simply a political document. It exists to provide political cover to members of Congress. But if that’s the case, it’s an unintentionally beautiful illustration of why industrial policy fails. It’s both economically crucial and very hard to allocate capital well; that’s why people who are good at it make so much money. Businesses struggle to do this well, and they’re really trying. What do you think the odds are that this is a wise use of money, when the people involved are barely pretending to try?

More here and here.