I was reading through comments at the WSJ the other day when a poster drew a connection between Cash for Clunkers and healthcare. The general idea followed this line of reasoning (paraphrasing):

Do we really want to entrust healthcare to the same entity that generated the massive failure of Cash for Clunkers.

I’m not really sure that is a fair question, as some I’ve asked have said that their opinion is that Cash for Clunkers was always meant to be a failure, financially at least. For some strange reason, that does almost make sense.

In any event, I was curious as to how Cash for Clunkers shaped up in terms of a return for the taxpayer.

I broke the analysis in the following parts:

1. Capital Outlay 1: total claims in dollars
2. Capital Outlay 2: financing expenses
3. Return 1: savings from the cost of oil
4. Return 2: dividends from auto manufacturers
5. Return 3: cash equivalent of the environmental impact

Each of these is viewed as marginal, in that they would not have otherwise occurred.

The executive summary: Cash for Clunkers created a -34% return for taxpayers.

A pdf showing the full analysis is attached here.

Essentially, the argument follows the line that tax dollars were used to stimulate the economy, aid the auto manufacturers, and generate a benefit to the environment. To accomplish this, we spent $2.9 billion. In return, we saved $1.3 billion on oil, and in theory would receive $1.1 billion in additional value or dividends through ownership of auto manufacturer stock (assuming all equity is owned by US taxpayers - probably not a valid assumption). We also saved CO2 emissions to the tune of 11 million tons. However, because not everyone can pay cash for these new vehicles, additional financing costs of $0.8 billion would need to be incurred.

The end result is that on our $2.9 billion investment, our return is ($1.0 billion).

To be fair, I tried to keep my assumptions, where used, conservative, in that I did not deliberately try to force this into a negative return situation. This statement alone will reveal my preconceptions, but if anything my awareness of these notions was heavily guarded against. From that standpoint, I think the bulk of the analysis can withstand argument. However, I would appreciate any feedback on the approach and results.

I return bearing gifts. After a short and well-used break from the site, I am glad to post a working version of a capital budgeting model. This model is straightforward, and focuses on corporate finance in terms of project analysis.

It incorporates all of the standard features, such as IRR and NPV calculations, but it adds a bit more. These special features include:

Scenario Analysis The user can compare multiple scenarios (3) side by side.
Sensitivity Analysis The user can manipulate inputs with slide bars to immediately see the impact on profitability.
SML Hurdle Rate The user can follow the SML methodology to develop a project specific hurdle rate.
Equivalent Annual Annuity The user can use the EAA method to compare projects with varying lives.
Multiple Depreciation Method The user can select straight line, or use a custom (commonly accelerated or MACRS) depreciation schedule.
Graphics The model automatically creates presentable 3D graphics to showcase the results based upon the selected inputs. These can quickly be copied and pasted into other MS software, such as PowerPoint.

As usual, the model has been uploaded to the Models page of the foquant.com site. The “Terms” tab has not been completed, but once uploaded that will share additional insight into the inputs and outputs of the model. Also, as usual, the user assumes any and all risk - I partake in model creation for recreation; if any reliance is placed upon these models it is at the assumption of risk of the user.

The Models page at foquant.com now features a binomial bond valuation tool. This is a crude first run at this, limiting the valuation to a bond with four periods to maturity.

I wanted to get this thrown together so I can expand upon the periods and the ability to value options on bonds. For now, though, the user can input their own rate tree, or choose a constant rate, and also dictate whether or not the bond has a coupon rate. The math that goes into the binomial tree is quite simple, so if a user needs to add additional periods it should not be exceptionally difficult.

Update 1: I have now also uploaded the same model but allowing for a callable bond with a user-defined call schedule.