Bipole III: Statistical Mutilation, Bad Math, and Outright Lies

So, in the never-ending bipole debate two numbers have recently been bandied about as the differences the cost of the east and west side route. The PC's are using $11,748, the NDP are using $13.68. Now when the $13.68 was first proposed (after the $11,748, had already been set forth) I figured someone more knowledgeable than me would come forward and destroy one (or both) of the numbers, because the completely absurd disparity between the numbers means someone (or multiple someones) is either lying or mutilating statistics to the point it might as well be a lie.

But other than some credulous partisanship, and a bunch of crap in the comments I can't say I've seen much talk on it. (The best discussion of the numbers I've seen are from the Purple Rod comments a few months back, before the $13.68 estimate). I have certainly not seen anybody analyze how two such vastly different numbers could possibly be put forward. (It is possible I may have missed it).

Time went on, then this from yesterday, and still no real analysis beyond a classic Abbott and Castello routine (of which I approve).

So given that nobody else seems to think these numbers are absurd enough to analyze, here goes:

First, the PC's number.

The cost difference between west side line vs. east side line is an estimated $3.62 billion. East Side: $788 million ($600 million for the line, plus $188 million for licensing) West Side: $4.1 billion plus $300 million line losses (the value of the energy lost in transit via a longer route).

A decent start, but perhaps somewhat unfair. The east side cost has almost doubled since the original cost, so it's quite possible the original cost of the west side was also underestimated. On the other hand, I know of no more recent estimate so I can't argue it's wrong.

Also, the $4.1 billion was an unofficial leaked figure and the NDP claimed that it was out-of-date and meaningless and the Hydro CEO says he never saw it. But then again the accepted NDP price estimate has increased from $2.2-billion to $3.2-billion, something they insisted on until very recently.

So take that as you will.

The only people who know the true projected costs work for Hydro and would likely be risking their job if they provided information not cleared by higher-ups.

By using a Manitoba population figure of 1,232,550, the per person differential cost is estimated at $2,937. This population figure is not a current Statistics Canada number. It looks like an about June 1, 2010 number.

Using the most current population number of April 1, 2011 of 1,246,396 gives a per capita value of $2,904.

The difference in population due to the date is a minor issue and not really worth mentioning (especially given that these are not census numbers and are likely closer to estimates than an actual head count).

The main problem is that the use of population is somewhat inaccurate. The Manitoba population will (hopefully) not be paying for this, ratepayers will. (Obviously there is a huge overlap, but the distinction between ratepayers and taxpayers is important). A better number would be 510,000, the number of Hydro ratepayers. This would increase this number particular significantly.

But then again, ratepayers includes businesses and industry who would use much more power than your average home owner. I found this on residential vs. commercial/industrial usage page 16 of this forecast from 2002, where it states that general usage (all commercial and industrial use) accounts for 65.9% of the total usage, according to page 10 33.7% is residential, while street light usage takes up the last tiny chunk. I can think of no reason why this would have changed significantly in the last decade. According to this survey from 2010, there were 439,096 residential customers (which would mean about 70,000 industrial customers).

But this distinction between residential and commercial might be meaningless. Any greater costs in rates accrued by a company would simply be passed on to consumers through higher prices and/or onto workers through lower wages, so the general Manitoba population would be paying for most of the "general use" rate increases anyways.

Continuing on:

So, if the differential cost is $2,937, then the differential cost for four individuals is 4 x $2,937 or $11,748.

In this analysis the number of families of size four is not used at all. It’s a per individual approach.

This is shoddy use of statistics. If they wanted cost per family, they should have calculated by household (which is available from StatsCan) not by individuals/4, but this is fairly irrelevant because the use of taxpayers rather than ratepayers is invalid on the face of it.

Conclusion: The PC's showed poor use of statistics and a gross error in the use of taxpayers.

So onto the NDP's number. (I know technically it's Brennan's number, but I have little doubt he was pressured by NDP to play along with their questioning for political purposes. At least, I really hope the CEO of Hydro is not actually as incompetent as this discussion makes him out to be).

So when I multiply that number by the number of families that the chief statistician tell us exist in Manitoba, I get a total amount of $3.8 billion. Now I’ve looked at the presentation you made. The total cost estimate of the bipole is $3.2 billion. So this advertising seems to suggest a total that’s more than the cost,

See above for earlier talk on the disagreement over the cost. I'm not sure which I'd use (so I'm actually going to calculate with both later on) because both are playing partisan games with numbers, but to attack the PC's calculations based on a potentially valid number from a leaked document strikes me as fundamentally dishonest. So, write off the bat we see the NDP are using a fundamentally different start point from the PC's.

yeah, I took the difference in length between the two routes. I took the total transmission cost, calculated the cost per kilometre, which really works out to quite an expensive amount I think it was $940,000 a kilometre, and applied that to the incremental length and got a number of $428 million,

This is just plain sloppy calculation and borderline dishonest. This assumes that the cost per kilometer on both sides is the same. Whether it is, I have no idea, but there is not way I would base a cost estimate on that assumption. There are far to many variables, such as terrain, existing infrastructure, political impediments, land use compensation, etc. to make it even remotely reasonable to make this assumption.

Second, his number of $940,000 is an outright lie (I really hope the CEO of Hydro is not this incompetent). The cost was stated as $3.2-billion; according to Hydro documents, the preferred route is about 1,364 km long. Simple division makes the cost $2,346,041/km. Almost exactly two-and-a-half times the amount he calculated.

The east side route is about 479 km shorter. The calculated extra cost on this is $1.1 billion, again about 2.5 times greater than Brennan's number.

I did not include what Mr. McFadyen was talking about, increases losses that occur, I excluded that. But that, it wouldn’t double this number, that’s for sure, it would be even less than that.

Notice this. I have no idea if it's double or not, but Brennan has stated that he didn't include part of the PC's calculations. The numbers are automatically not comparable because of this. Any number Brennan gives after this will automatically be lower than the PC's number. Any claim the numbers are comparable is a straight-out lie.

So then I took the number of households, escalated up to 2017. And that number’s less than the current number of customers, and I took the incremental cost per household (inaudible…)

That number came out to $821 per household.

At least he's using households not individuals, but he should be using ratepayers, not households. Similar problem as with the PC's number. The last census counted 448,780 households in Manitoba in 2006. So, $953/household if we use Brennan's earlier number. $821 sounds reasonable given that the number of households has increased since 2006. But his math was wrong earlier, so the number is actually almost $2,500 per household, but we can say $2,000 as the number of households has increased since 2006.

And then I said, well, that’s over the life of the line so I divided that by 60. And so the annual cost would be $13.68.

After all the bad math and poor statistics, this one is the most disgustingly dishonest. You can not compare a total number like the PC's with an average yearly cost over 60 years like this one. The dishonesty of it just reeks.

Not to mention that if we amortize the cost over 60 years like this calculation suggests, the interest rates will well over double the cost of the project. You can play with this amortization here, but even if we put everything in Hydro's favour: 1 payment a year, a low 3% interest rate, the total interest paid will still be 140% of the cost. So, interest costs would be more than the project itself. (I really, really hope the CEO of Hydro is not incompetent enough to amortize over 60 years).

If we aren't amortizing over 60 years, then the rate increases will not be paid over 60 years, they will be paid over the period of amortization. So the use of 60 years is fundamentally dishonest.

Conclusion: The NDP/Brennan abused statistics and math to the point where I consider it an outright lie and they even got basic division wrong. The PC's calculations were sloppy and based on poor assumptions, but the NDP/Brennan calculations were just disgustingly dishonest.

Now onto a more accurate number than either of those sad, partisan little bits of statistical mutilation.

I'll calculate a few numbers using both a low using the NDP's $3.2-billion and a high using the leaked report's $4.1-billion.

The average cost per ratepayer (510,000) for constructing the entire west side line could range from $6275 to $8039.

The average cost per ratepayer (510,000) of the east side line using the $788-million estimate is $1545.

So, if we use the original estimate for the east-side line the additional average construction cost per ratepayer of constructing the west line over the east line could range from $4370 to $6494.

An honest assessment on the available public information would put the extra cost of construction (not including transmission losses) of the west side line at somewhere from $4370 to $6494 per ratepayer. These are the numbers I would use depending on which west side project cost estimate I accepted as true.

I'll perform some other calculations and why I wouldn't use them.

Extra total cost of west side project based on the average cost per a kilometer of transmission line (calculated earlier): $1.1 billion.

Average cost per ratepayer (510,000) based on this: $2157.

This number is not exactly dishonest, but, as mentioned earlier, it assumes that the cost per kilometer is the same in both projects, which is not an assumption I would make unless there was some hard proof. It's use is fundamentally unsound unless said hard proof is offered.

Now let's calculate the cost per residential ratepayer, this will be a bit more complex, so I'll put the math in.

$3,200,000,000-788,000,000 = $2,412,000,000
(Extra total cost of west side line).

$2,412,000,000 * 33.7% = $812,844,000
(Cost of west side line borne by residential ratepayers).

$812,844,000 / 439,096 = $1851
(Lower limit extra cost of west side line per residential ratepayer).

$4,100,000,000 - 788,000,000 = $3,312,000,000
$3,312,000,000 * 33.7% = $1,116,144,000
$1,116,144,000 / 439,096 = $2542

So, the total extra construction cost (not including transmission losses) of the west side line per residential ratepayer would be somewhere between $1851 and $2542.

This would be a perfectly legitimate number to use. I would not use it myself, because the extra costs borne for commercial and industrial use will simply be passed on to either Manitoba consumers or workers as mentioned earlier. But let's calculate commercial and industrial ratepayers increased costs.

$2,412,000,000 * 65.9% = $1,589,508,000
$1,589,508,000 / 70,000 = $22,707

$3,312,000,000 * 65.9% = $2,182,608,000
$2,182,608,000 / 70,000 = $31,180

So, the total extra construction cost (not including transmission losses) of the west side line per a commercial/industrial ratepayer would be somewhere between $22,000 and $31,000.

Transmission losses have been estimated at about $300-million by a group of engineers. So that's about an extra $588 per ratepayer of losses.

The costs of this will borne by the Manitoba consumer, so prices of goods and services will go up (or it will be borne by the workers, in which case wages or benefits will go down).

So there you have it, the actual extra costs of the west side route as opposed to the east side route calculated as accurately as possible with the information publicly available.

As for what side of the debate I support, I support the East Side route because the experts that are not being leaned on by the government to express a certain view support it, and being no engineer and seeing no real flaw in their arguments I accept their expertise.

BTW, their estimate was $4,200 for a family of 5, or about $840/person. Although, it is not clear if that includes transmission losses or not. They are also using old estimates from almost a year ago, and the cost of the west side has increased.