Monday 21 May 2012

BERL and asset sales

The Greens commissioned BERL to look at the government's planned set of asset sales.

Recall that I've previously argued that the "but the bond financing cost is lower than the flow of dividends" argument is nonsense because it says the government should borrow to invest in the stock market where stock returns are higher than what the government pays in interest; it ignores that stocks are riskier assets than New Zealand government bonds.

What really matters is whether an asset is better managed publicly or privately. If the assets are more efficiently held publicly, the "loss of flow of dividends" critique can make sense: in that case, a private owner is willing to pay less for the shares than the flow of dividends is worth to the government. Otherwise, a high dividend flow just means the asset's selling price is bid up. If the private owner expects efficiency gains, competitive IPO markets push the asset's selling price to being higher than the discounted value of the current revenue stream.

So, how does BERL approach the problem? They assume that revenues from asset sales are used to build other assets that yield dividends equal to the returns on the sold assets but that time-to-build means a few years' delay in getting the flow of assets from the alternative stream. It's then not particularly surprising that they find that asset sales are a dumb idea. It would be hard to find anything other than "privatization is a dumb idea" given that starting point. They also assume that borrowing costs are lower than dividend yields and conclude that it makes more sense to borrow than to sell off assets. They do some year by year projections going forward on the basis of the assumptions, but all that time path depends on the question-begging at the outset; I'm not going to get into whether they got the time series right.

BERL also seems pretty worried about effects on the country's net external debt position. So they set up scenarios comparing asset sales, where buyers may be foreign or domestic, with a bond issue, where bonds are assumed to be bought only by domestic investors. On this basis, they find that the asset sales will hurt net foreign liabilities. Perhaps their conclusion would have changed if they considered that foreign investors do sometimes also buy our government's debt, or that domestic investors can on-sell government bonds to foreigners.

Further, when BERL makes the case for debt over asset sales based on the difference between the government's cost of borrowing and the dividend yield from state owned enterprises, they don't seem to adjust for that dividend yields tend to be higher because asset owners need a risk-based return. If it doesn't make sense to take out a mortgage on your house at 5% because you can buy stock in a company that usually pays 6% dividends, it probably doesn't make sense for the government to do it either.

As a fun robustness check, they compare their results against a scenario where the new investments yield lower dividends than the new investments. Unsurprisingly, they find that privatization is then even worse!

I'll agree with BERL that some of the benefits of partial privatization seem overwrought. I've been critical of partial privatization, and especially of starting with the energy companies. But if this is the best case against partial privatization that the Greens can come up with, it sure isn't convincing.

Previously:

15 comments:

  1. Do you have to pay BERL extra to get them to work backwards from the appropriate conclusion to the necessary assumptions and does the extra payment increase along with the outlandishness of these assumptions?

    ReplyDelete
    Replies
    1. You'd have to get in touch with them for an answer on that one; I've never tried hiring BERL to do anything. I have, on the other hand, probably sent a fair bit of work their way. There's a big market for their kind of product, and it's now pretty common knowledge which firm reliably delivers it. Where the ultimate audience is somebody catching a 3 second clip on 3News, "sciency" is good enough.

      Delete
    2. "Do you have to pay BERL extra to get them to work backwards from the appropriate conclusion to the necessary assumptions"

      Remove the word "extra" and I think you'll be closer to the mark.

      Delete
    3. Most cost benefit analyses in government (Right and Left) are instrumental ie. they start with the desired conclusion, then set the variables to achieve a safe NPV or BC ratio, then find the evidence to support the assumptions/variables. By very skeptical.

      Delete
    4. Have a look at the Taylor Duignan Barry one that predicted cost savings resulting from the supercity transition. Have your rates gone down yet....?

      Delete
    5. We need some more effective mechanism for shaming the worst providers of these things...

      Delete
  2. I'm ok with those assumptions. One of the benefits of asset sales, as listed in the Budget Policy Statement 2012 is to "Free up capital for the Government to invest in other public assets", as noted in page 11 of the PDF. Further, if the new assets do not yield dividends equal to or greater than the existing assets then that just strengthens the case against asset sales, so using that assumption as a basis for the report seems charitable to me.

    ReplyDelete
    Replies
    1. Except that it's spurious, like most of the reasons that the Government has put forward for the asset sales.* If these "other assets" have a positive risk-adjusted return then we should invest in them regardless of how they're funded. If they don't, we shouldn't. The only way this would make sense is if the Government were funding-constrained, which isn't the case at the moment although I guess there's a risk of that in the future if Europe continues to circle the drain.

      (* As in, they've focused on touchy-feely arguments like "we'll spend it on schools!" rather than a doomed effort to sell the public on the economic benefits.)

      Delete
    2. "Dividends" is a bit of a term of art here. The SOEs pay dividends, but without a market test on valuation, it's a bit hard to say what the rate of return really is. Some of the investments National's considering are schools - they do good stuff, but hard to quantify a return. Even still, have the assets that National's proposed partially privatising given anywhere near an 8% return recently?

      Delete
    3. @Miguel: I'm wondering whether Treasury's starting to be worried about being funding constrained. But agreed.

      Delete
    4. Possibly. But the pre-election fiscal update last year, when Europe was already starting to blow up, was based on ten-year borrowing rates sedately rising from 4.4% to 5.4% over the course of five years. I'll make a friendly wager with you that the projections in Thursday's Budget will be lower than that.

      Delete
    5. @Miguel: That narrows the gap even further then.

      Delete
    6. COMU (an arm of Treasury) reported in their 2011 Annual Portfolio Report that the four energy state-owned assets up for sale made an average total shareholder return of 18.5% pa over the last five years. 'Total shareholder return' measures performance from an investor perspective and includes the combined return from dividends and growth in the value of the company. Treasury consider this the best performance measure to use from the Government's perspective.

      Put simply, our energy SOEs are cash cows.

      Delete
    7. I'd heard lower numbers; am chasing after them now.

      We can pretty easily determine the revenue flows from SOEs, and we can pretty easily determine their expenditures. Adequately valuing the asset base is a fair bit tougher though. Thinking about Meridian, for example, their assets aren't just the dams they have; it's also the regulatory rent they're drawing by virtue of having had permission to build dams before we banned building new dams. That regulatory asset is every bit as real as the concrete. Surely we ought to be imputing a fair bit of the excess of their current revenue over their current costs as being a normal return to a capitalised regulatory asset; whoever would bid to buy Meridian would be willing to pay extra because of the existence of that regulatory asset. No?

      Delete
    8. @Anon: I was looking at dividend rates. Total shareholder returns include increases in share prices, but if the shares are never sold, it isn't like the government can choose to fund new alternative investments out of the total shareholder return figures - unless it starts borrowing against those shares.

      Why wouldn't we expect the sale price of cash cows to be bid up to the point where owners expect only a normal risk-adjusted return? My worry on this front is that in a political second best, the power companies really are better owned by government: high profits going to private power companies in dry years would attract price controls. But that's straying pretty far from the kind of argument BERL was advancing.

      Delete