America’s Underinvestment in Public Infrastructure

Real per Capita Public Investment vs. GDP, 1950-2013

Public infrastructure in the United States is an embarrassment.  This is clear even to ordinary travelers.  Countries in Europe and in much of East Asia enjoy far better roads, highways, public transit, and other forms of public infrastructure than the US does, even though the real per capita incomes of these countries are lower than that of the US.  And this is backed up by more systematic global comparisons, such as in the Global Competitiveness Report of the World Economic Forum.  The most recent report ranked the US as only number 15 in the world in terms of its infrastructure (transport, power, and telecom).  This put the US behind Canada, the major countries of Western Europe, and such countries as Japan, Korea, Hong Kong, Singapore, and Taiwan in East Asia.

The poor quality of public infrastructure in the US should not, however, be a surprise.  As the chart at the top of this post shows, the US is spending no more now, in real per capita terms, than it did over a half century ago in 1960, in the last year of the Eisenhower administration.  The chart draws on data issued in the standard GDP (NIPA) accounts of the BEA of the US Department of Commerce.  Infrastructure investment is taken to be total government investment (at all levels of government – Federal, State, and Local) in structures, excluding such spending by the military.  Most government infrastructure spending in the US is for transport (primarily roads and associated bridges, but also including investment in mass transit, ports, and airports), with a significant amount also for water and wastewater treatment.

Public infrastructure spending in real per capita terms rose during the Eisenhower administration in the 1950s (when the Interstate Highway system was started) and continued rising during the Kennedy and Johnson administrations in the 1960s.  Indeed, during this period, such spending rose at a somewhat faster pace than real per capita GDP, the blue line in the chart.  But starting in 1969, the year Nixon took office, public infrastructure spending was cut.  By the mid-1970s it was down close to the level seen at the end of the Eisenhower administration (in real per capita terms), and then was cut even further at the start of the Reagan administration.  It then began to increase from 1984 with this continuing to a peak in 2002, after which it fell again.  By 2013 it was 2% lower than it was in 1960.  Over this same period, real per capita GDP almost tripled.

In dollar terms, real per capita spending on public infrastructure (in terms of 2009 prices, the base now used in the GDP accounts) was $793 in 1960 and was 2% lower, at $776, in 2013 (about 1.6% of GDP).  Over this same period, per capita real GDP rose from $17,159 in 1960 to $49,852 in 2013.  The increment in real per capita GDP was $32,693 over this period.  None of this growth went to increased investment in public infrastructure.

It is this stagnation in real per capita spending, and huge lag behind income growth, that has led to bridges and highways that are both congested and in poor condition.  People drive more, fly more, and import and export more goods, as their real incomes grow.  Public infrastructure has not kept up.  A 2009 report issued by the American Association of State Highway and Transportation Officials (AASTO) notes that vehicle miles driven between 1990 and 2007 rose by 41%, about double the increase in the US population over this 17-year period (of 20.6%).  Based on the figures in the chart above (which however covers all public infrastructure, not just highways), spending to build or maintain such infrastructure per mile driven fell by over 20% over that period.

The AASTO report also found (based on an analysis of US Federal Highway Administration data) that one-third (33%) of the nation’s major highways was classified as being only in poor or mediocre condition (as of 2007).   Thirteen percent was classified to be in poor condition, with this rising to over 60% poor in some major urban areas.   And roads in poor or mediocre condition deteriorate quickly, leading to much higher costs when the road eventually has to be repaired.  The AASTO report notes that the cost per mile over 25 years is three times higher if roads are left to be reconstructed, instead of maintained on the regular recommended schedule.

This stagnation in real per capita spending on public infrastructure over more than a half century may be surprising to some.  While many might be aware that infrastructure spending has not kept up with real per capita GDP (which has almost tripled), most people would assume that there has been at least some increase in per capita infrastructure spending.  But that is not the case.

Part of the reason for this mis-conception is that when measured as a share of GDP, it might not appear that public infrastructure spending has fallen so far behind.  As a share of GDP, public infrastructure spending (using the figures cited above for public investment in non-Defense structures, from the BEA accounts) was 39% less in 2013 than it was in 1960.  Put another way, public infrastructure spending would have had to increase by 64% (=1/(1-.39)) between 1960 and 2013, to match the GDP share it had in 1960.  But the figures shown above in the chart indicate that public infrastructure spending would have had to triple over this period to match the increase in GDP.

Why this big difference?  The reason is Baumol’s Cost Disease, which was discussed in an earlier post on this blog.  If the price index for public infrastructure spending over this period had matched the price index for overall GDP, then an increase in infrastructure spending of 64% would suffice to bring it into line with the increase in real GDP over the period.  But the cost of building infrastructure has risen at a faster pace than the cost of making goods generally.  This is not because of increased waste, but rather because building infrastructure is by nature labor intensive and hard to automate.  The relative cost of infrastructure will therefore increase over time relative to the cost of goods whose production can be increasingly automated.

The importance of this is huge, but is often ignored in the debates.  As the chart above shows, investment in public infrastructure has stagnated in real per capita terms over more than a half century, and would need to almost triple at this point to catch up with how much real per capita GDP has grown.  This is far greater than the 64% increase (which is itself not small) that one might assume would be necessary by simply focussing on GDP shares.

The fundamental cause of this stagnation in real spending on public infrastructure has been an unwillingness in Congress to pay for it.  The most important source of funding for highway expenditures has been the gasoline tax, which supports the Highway Trust Fund. But as was discussed in an earlier post on this blog, gasoline taxes have been set as so many cents per gallon and are not adjusted regularly for inflation.  The last time the tax was raised (in nominal terms) was in 1993, over 20 years ago.  Since then, even general inflation has eroded this by over 50%.  If one took into account that prices for infrastructure investments rise at a substantially faster pace than general prices (due to Baumol’s Cost Disease, discussed above), the real erosion has been much greater.  As a result, funds in the Highway Trust Fund are far from adequate.

The result has been repeated crises as the Congress passes one short term patch after another to allow even the overly low on-going highway investments to continue.  One such crisis is underway now, where expenditures would need to be slashed on August 1 if nothing is done.  The Senate is currently expected to vote this week on an extension, although it would only be for a few months at best.  If passed and can then be reconciled with a similar House passed measure (passed two weeks ago), spending on highway investment will be able to continue for a few more months.

To provide the needed funds, given that the Highway Trust Fund is far from sufficient (due to the failure to adjust the tax to reflect inflation), Congress has included again an especially stupid provision in the draft bills.  As it did in an earlier authorization in 2012 (see the blog post cited above), Congress would allow corporations to make assumptions on their pension obligations which will in effect allow them to underfund their pension obligations by even more than currently.  The corporations will then show (on their balance sheets) higher profits, which will generate somewhat higher corporate income tax obligations.  These higher tax obligations will be counted as government revenues.  But those reliant on corporate pensions will be at greater risk of not receiving the pensions they are owed.  Ultimately the government may be obliged to cover these pension obligations (through the Pension Benefit Guarantee Corporation).  But these costs latter costs are being ignored.