Tuesday, September 18, 2007

LSIF critic spins tales for Fraser, Howe instututes

The Fraser Institute just published a report on labour-sposored investment funds (LSIFs) that is almost identical to a report published earlier this year by the C.D. Howe Institute. And with good reason: both papers have the same lead author -- Douglas Cumming.

Cumming doesn't think much of LSIFs and thinks governments in Canada at provincial and federal levels should discontinue the tax credits provided to LSIF investors. Ontario has already announced that it is phasing out the LSIF tax credit.

There are legitimate criticisms that can be made of LSIFs, but for years Cumming has been augmenting these bona fide critiques with ideologically-driven spin on what is often dubious data to begin with.

Reports from Howe and Fraser are usually so driven by ideology that you rarely have to read past the title to know what they're going to say. That's particularly true for these reports, since Cumming has been distributing different versions of the same report for the last seven years.

The Canadian VC industry has seen a lot of changes over that time, but as an ideologue trying to support his views, Cumming is always able to draw the same conclusions. To use a Stephen Colbert line, "he believes the same thing Wednesday that he believed on Monday, no matter what happened Tuesday." And, in this case, Tuesday is the last seven years, and a lot has happened, but Cumming has been a rock of consistency throughout these wildly-changing times.

Which isn't all that surprising, seeing how stale much of his data is. He has claimed for years that LSIFs "crowd out" other VC funds from the market. The most recent data he uses in this analysis is now six years old. He actually uses data from more years in the 1970s than in the 2000s.

While Cumming will occasionally mention techniques like regression analysis to provide a veneer of legitimacy, his favourite tool is spin. For example, Cumming says that LSIF managers -- on average -- have more than twice the number of companies in their portfolio as VC managers. The conclusion he chooses to draw from this is that LSIF managers are just too swamped to become involved in meaningful ways with their companies and, therefore, fall behind on the learning curve compared to VC managers.

But this is one of those great "facts" (and it's questionable whether it's a fact at all, even though it would be fairly easy to check) that can be used to support whatever you want it to. If you're an LSIF supporter, you can conclude that LSIF managers, by being involved with more companies, get to see more situations, interact with more entrepreneurs, and deal with more business issues in a shorter period of time than their VC counterparts. That would put them ahead on the learning curve. It's not any more reasonable than Cumming's position, but neither is it any less-supported by the data he cites.

Another example: Cumming says that LSIFs don't have to be as concerned about ROI as other VCs (because investors get the tax credits no matter how the fund performs), and this enables them to spend more to attract deals in a non-level playing field where VCs can't easily compete. Okay, but if it's true that LSIFs can spend more, then -- for the same reasons -- they should be able to outbid VC firms for management talent. That would mean that LSIFs should be able to attract a more talented team than VCs. But that's not a conclusion Cumming wants to promote, so he doesn't mention it -- even though it's as much a consequence of his assumptions as the point of view that he does want to support.

One more, and then I'll stop. Cumming points out that individual LSIF investors only account for a tiny percentage of the amounts put into any fund, as opposed to a limited partner of a VC fund, who would account for a far higher percentage of the total fund. The conclusion he wants to promote is that LSIF investors have much less control over fund management than VC investors and exert less pressure on managers to perform. And that's reasonable enough, but an LSIF supporter could take that very same data, and point out that LSIFs have to raise money on a continuing basis (or, more realistically, an an annual basis when RRSP season comes around) and, therefore, have to face the market every year and convince them to invest. The pressure to perform would be continual. And LSIF investors typically commit for a slightly shorter period than LPs with VCs, which again makes it easier for them to walk away from a poor performer. I don't buy that argument either, but it's as reasonable as Cumming's.

That's the great thing about "data" -- with a little selectivity and spin, you can reach just about any plausible conclusion you want ... as long as you're not too thorough or rigorous.

I'll continue with this topic -- and the LSIF industry's response to Cumming (also flawed) -- in a future post.