Navarro Nonsense on Privatization of Public Infrastructure

Brad Plumer has a must read discussion on Trump’s plans for infrastructure:

What Trump has right now is an idiosyncratic proposal for Congress to offer some $137 billion in tax breaks to private investors who want to finance toll roads, toll bridges, or other projects that generate their own revenue streams. But this private financing scheme, experts across the political spectrum say, wouldn’t address many of America’s most pressing infrastructure needs — like repairing existing roads or replacing leaky water mains in poorer communities like Flint. It’s a narrow, inadequate policy. For instance: “This is unlikely to do much for road and bridge maintenance,” notes Harvard economist Edward Glaeser. “And [economists] have long believed that the highest returns are for fixing existing infrastructure.”

This proposal was written by Peter Navarro and Wilbur Ross. No kidding. As I noted even Greg Mankiw stated how their analysis of the macroeconomics behind the Trump tax cuts was nonsense. But the decision to publicly build infrastructure versus having the private sector has elements of financial economics embedded the discussion so maybe these two have some insights. Navarro and Ross are correct that we are under investing in infrastructure but their political spin on this is beyond belief:

For example, with Hillary Clinton’s full support, only 5% of Obama’s $840 billion program of infrastructure spending, initiated in 2008 at the depths of the Great Recession, was actually spent on “shovel ready” projects. The rest was dissipated, with little stimulus result while our nation’s infrastructure gap has widened.

Never mind the fact that Obama’s advisers wanted more infrastructure investment and less in the way of tax cuts but Obama had to deal with Republicans who wanted all tax cuts all the time. But this is not my problem with their paper. I will admit that my two cents on this was not a real analysis:

I’m not sure how the city decided that $1150 million was fair market value but let’s do a small DCF model that starts with nominal profits at $58 million per year but let’s this figure rise by 2 percent per year. The fair market value depends of course on what we assume the appropriate nominal discount rate should be. If one is willing to assume a 6.9 percent nominal rate, then this was a fair deal to the taxpayers. The current interest rate on 30-year government bonds, however, is only 3 percent. If we use this as our discount rate, the fair market value would be over $3 billion. Of course this has been an old story – government officials selling taxpayer assets to private companies at bargain prices.

I guess one could argue that the appropriate discount rate should include both the risk-free rate (which I took to be 3 percent) but also a small premium for bearing systematic risk. But I found the Navarro-Ross discussion of the cost of capital issue to be incoherent. Let’s quote the relevant portion of this discussion in full:

For infrastructure construction to be financeable privately, it needs a revenue stream from which to pay operating costs, the interest and principal on the debt, and the dividends on the equity. The difficulty with forecasting that revenue stream arises from trying to determine what the pricing, utilization rates, and operating costs will be over the decades. Therefore, an equity cushion to absorb such risk is required by lenders. The size of the required equity cushion will of course vary with the riskiness of the project. However, we are assuming that, on average, prudent leverage will be about five times equity. Therefore, financing a trillion dollars of infrastructure would necessitate an equity investment of $167 billion, obviously a daunting sum. We also assume that the interest rate in today’s markets will be 4.5% to 5.0% with constant total monthly payments of principal and interest over a 20- to 30-year period. The equity will require a payment stream equivalent to as much as a 9% to 10% rate of return over the same time periods. To encourage investors to commit such large amounts, and to reduce the cost of the financing, government would provide a tax credit equal to 82% of the equity amount. This would lower the cost of financing the project by 18% to 20% for two reasons. First, the tax credit reduces the total amount of investor financing by 13.7%, that is, by 82% of 16.7%. The elegance of the tax credit is that the full amount of the equity investment remains as a cushion beneath the debt, but from the investor point of view, 82 percent of the commitment has been returned. This means that the investor will not require a rate of return on the tax credited capital. Equity is the most expensive part of the financing; it requires twice as high a return as the debt portion, 9 to 10% as compared to 4.5 to 5.0%. Therefore, the 13 percent effective reduction in the amount of financing actually reduces the total cost of financing by 18 to 20 percent. By effectively reducing the equity component through the tax credit, this similarly reduces the revenues needed to service the financing and thereby improves the project’s feasibility.

Is this 5% the expected return to assets while the 10% is the expected return to equity? It seems that they are assuming that 83.3% of any project will be financed by debt while 16.7% will be assumed by equity but their writing is a bit vague where. I’m trying to also understand how their example is consistent with the standard Capital Asset Pricing Model as well as the Modigliani-Miller proposition but given their incomplete discussion, we need to do some assumptions. I guess they are assuming a 4% risk-free rate plus a premium of 1% for bearing operational risk. Given our assumptions about debt versus equity, this translates into a 6% premium for bearing both operational and leverage risk. Maybe these assumptions with respect to systematic risk are defensible but they have neither made the case nor even articulated that this is what they are actually assuming. Of course, a 4% risk-free interest rate strikes me as a bit high. There is so much more that I find questionable about their discussion but let’s stop there by making a simple request. Could they actually write a coherent discussion with respect to the cost of capital?