The IMF’s latest World Economic Outlook, in an unholy alliance with Larry Summers, claimed this week that a surge in publicly funded infrastructure spending would provide an increase in economic output with no downside risk. The truth is almost precisely the opposite: by indulging in ill-thought-out and boondoggle-filled public infrastructure spending, governments in rich countries leave their long-suffering taxpayers gasping for their next dinner.

That’s not to say that infrastructure investment is never productive. The Liverpool-Manchester railway, opening in September 1830, was unquestionably infrastructure, but it was built and operated by the private sector. As the world’s first passenger railway, it was of course highly profitable, both financially for its owners and economically in terms of its effect on the global economy as a whole.

Even this magnificent human achievement was not without its economic costs, however. At its opening on 15th September, 1830, a design flaw became apparent in that the two sets of tracks had been laid too close together, which resulted in William Huskisson, MP, standing by the Duke of Wellington’s carriage, being killed by a train approaching on the other track. The political and economic effects of this were unquantifiable but immense as Huskisson had been in the process of negotiating an agreement with Wellington, then Prime Minister, for his group of moderate MPs to join Wellington’s government, to strengthen it after setbacks in a recent election.

Without Huskisson, the group’s new leader, Lord Palmerston, did a deal with the opposition Whigs. As a result Wellington’s government fell, and the country was subjected to the Whigs’ Reform Bill, destroying the existing constitutional balance, followed by several decades of their inept economic management which eroded the economic pre-eminence that Britain enjoyed in 1830. Set against those costs, even the Liverpool and Manchester Railway, considered on a stand-alone basis without adding in the value of its example for others, looks a thoroughly dubious proposition. (On the other hand Huskisson’s death was reported around the world, apparently performing a highly beneficial viral marketing function for the new transportation system.)

The IMF’s plea relates however not to technologically innovative private-sector initiatives such as the Liverpool and Manchester Railway, but to government infrastructure spending in general, the great majority of which involves no technological innovation and no potential for superior economic returns (because otherwise the private sector would have built it already.) It claims that such projects generally bring positive net economic returns, greater in developed countries than in emerging markets and that now, while interest rates are exceptionally low, would be a good time for an infrastructure spending surge.

Larry Summers in the Financial Times last Monday, reinforced this argument, claiming that infrastructure spending is a “free lunch” and that the IMF study shows $1 of infrastructure spending increases output by $3 (in fact that’s the output from their model – Keynesian of course – not from their empirical examination of the question, in which the output elasticity of infrastructure spending is more like $1.30 per $1.) He also claims that infrastructure spending that yielded a 6% real return would produce sufficient tax revenue to pay for its funding cost, given that real interest rates are currently only 1%.

The problem is that very few infrastructure projects selected by bureaucrats yield anything like a 6% real return. The $80 billion California High Speed Rail project, the 43 billion pound (costed by the Institute of Economic Affairs at 80 billion pounds ($128 billion)) British HSR 2 project and the $150 billion 30-year Amtrak project for high-speed rail on the U.S. East Coast would none of them yield anything like a 6% real return, if indeed they were profitable at all.

Indeed the IMF undercuts its analysis by noting, buried in the small print on page 83, that its supposed economic of infrastructure spending depends crucially on how it is financed: “Although a debt-financed public investment shock of 1 percentage point of GDP increases the level of output by about 0.9 percent in the same year and by 2.9 percent four years after the shock, the short- and medium- term output effects of a budget-neutral public investment shock are not statistically significantly different from zero.”

To those of us with experience of GDP accounting, the IMF’s case then collapses. If infrastructure spending is not beneficial if financed by spending cuts elsewhere in the budget or by tax increases, but only if it is financed by debt, then the benefits of infrastructure come down to the Keynesian fillip from taking on more debt. As discussed in this column several times, GDP statistics were designed by 1930s Keynesian economist Simon Kuznets to produce an apparent increase in national output every time public spending was increased, regardless of whether that spending had any useful purpose. Thus the New Deal was justified.

In that sense, if you believe GDP statistics, the Keynesian promise of the expansionary effect of a useless public spending boost is self-fulfilling; GDP automatically increases by the amount of the boost. To measure the true effect of a public spending boost, or of any other public policy change, you have to look at Gross Private Product, GDP with the government left out, to determine whether the real productive economy has benefited or been harmed by the policy change.

In general public spending boosts, on infrastructure or anything else, don’t provide such a benefit, although clearly the key elements of say the Interstate Highway System may be genuinely stimulative. (It is however suggestive that the Eisenhower years, when that system was begun, were ones of very sluggish GDP growth, with two recessions in three years in 1958-60, when Interstate Highway spending was expanding most rapidly.)

Thus the increase in output from a boost in debt-financed infrastructure spending is purely an artifact of Keynesian national income accounting, and does not occur if the infrastructure spending simply replaces other public spending. As far as the private sector is concerned, the average benefit of an infrastructure program boost is negligible.

There is however a further cost of debt-financed infrastructure, not captured in Keynesian economic models, which is the effect on the country’s future fiscal flexibility of the additional debt. For a private company, the cost of capital for a boost in spending financed by debt would not be the cost of debt alone, but a blended mix of the cost of debt and the cost of equity, with the debt increase itself increasing the company’s leverage, hence increasing the risk and assumed cost of its equity.

A similar calculation should be adopted in relation to government financing. Japan’s 1990s infrastructure mania appeared to be fairly cheap when considered against the very low yen interest rates then prevailing. However, it left the country with a permanent budget deficit that has transformed its fiscal situation from a healthy low-debt position (Japan’s public debt was about 60% of GDP in 1990) to a position teetering on the edge of bankruptcy, with public debt of 240% of GDP, close to the highest level ever survived without outright default.

When calculating the cost of additional infrastructure, in a country whose infrastructure was already amongst the world’s most extensive, Japanese bureaucrats should have factored in the cost of wrecking their country’s fiscal position. That should have made them step back from the unnecessary additional projects, however beneficial they were to the country’s Keynesian national accounts and the LDP’s short-term electoral prospects. Junichiro Koizumi spotted this problem, during his term of office in 2001-06, and cut back on infrastructure spending, moving towards a balanced budget, and seeing the real economy boom, but alas his term in office was all too short at five years.

There is a further problem with a boost in infrastructure spending, and that’s the cost of the infrastructure. As the high-speed rail projects discussed above suggest, this has escalated out of all measure in the last 50 years, as a result of bureaucratic delays, union featherbedding and legal harassments. Even for the new technology of high-speed rail, a comparison with French project costs in the 1980s shows that modern costings are way out of whack, inflated three or four times above their proper level. Go back further, and compare the $10 billion cost of the proposed but currently abortive Hudson River road tunnel with the $700 million (in today’s money) cost of the very similar Holland Tunnel, completed in 1927, and you see that infrastructure cost inflation has got completely out of hand. Until those costs have been wrung out of the system, no new major infrastructure projects should be started; they are simply at current costs unaffordable.

In today’s circumstances, the world’s central banks can print money all they want. They simply depress the real return on savings below zero, and de-capitalize their countries’ economies. A boost in spending on cost-bloated infrastructure, whether financed by debt or by money printing, would simply deplete the spending countries’ private sector capital bases still further, and depress their citizens’ long-term living standards. Far from being desirable, as Summers and the IMF shrilly claim, it would be thoroughly damaging to our welfare.

Much better to suffer a few potholes than to impoverish ourselves further!

-0-

(The Bear’s Lair is a weekly column that is intended to appear each Monday, an appropriately gloomy day of the week. Its rationale is that the proportion of “sell” recommendations put out by Wall Street houses remains far below that of “buy” recommendations. Accordingly, investors have an excess of positive information and very little negative information. The column thus takes the ursine view of life and the market, in the hope that it may be usefully different from what investors see elsewhere.)