A New Meaning for S&L

It’s a commonplace that the Cold War had a great deal to do with shaping the global economy in the second half of the 20th century.

The rivalry between the Soviet Union and the United State started with the atom bombs, long-range bombers and radar. The doctrine of massive retaliation was designed to spare each country the need to prepare for every possible contingency. It was enough to deter the foe by threatening that neither side could emerge the winner in a nuclear exchange.

Thus powerful military-industrial complexes in both countries grew up in place of enormous standing armies. A steady stream of government-funded R&D produced a series of rapid technological developments — notably rocket science (meaning control techniques), nuclear power and the digital computer.

Then all that rivalry ended suddenly with the fall of the Berlin Wall and the subsequent collapse of the Soviet Union. To be sure, US spending on big weapons systems has continued a fairly high rate for a decade, mainly out of political momentum. It is remains significant, but at a lower level than when the arms race was underway.

So what changes, if any, should the US make in defense R&D policy in the anti-terrorist Era? With what anticipated economic effects?

A pretty good working hypothesis has been ventured by Manuel Trajtenberg, an economist at Tel Aviv University. Trajtenberg’s credentials for thinking about this sort of thing are pretty good. A native Argentinean, he earned his Harvard PhD in 1984, then spent five years writing a landmark study of CAT scanners as exemplars of the economics of product innovation.

In 1995, he and Stanford University economist Timothy Bresnahan introduced an influential model of “general purpose technologies” as engines of growth — the steam engine, for example, compared to electricity, to the laser, to the computer. He has advised the Israeli government on innovation policy, serves on corporate boards, and otherwise keeps abreast of a wide swathe of emerging technologies.

He presented “Crafting Defense R&D in the Anti-terrorist Era” last week to a meeting of science policy experts organized by the National Bureau of Economic Research in Washington, D.C. To be published the autumn as the lead article in the fourth volume of an NBER series on Innovation Policy and the Economy, the paper and the model on which it is based are expected to be posted meanwhile as a working paper of the Foerdor Institute.

First off, Trajtenberg asks, has anything really changed? Won’t China be along shortly to challenge the technological superiority of the US military, at least in nuclear warfare? (Remember the panic over the supposed infidelity of US government scientist Wen Ho Lee?) Not if the stock of knowledge has anything to do with it.

Computed by adding up lagged R&D expenditures and depreciating them at an annual rate of 15 percent — much faster than the stock of physical capital — the knowledge stock is shown by such calculations to be much larger than that of any other nation in the world. Only the Soviet Union came close, and its knowledge base has shrunk dramatically since its collapse. Russia cannot afford to renew it.

And as for China, Trajtenberg consider a serious challenge is “rather unlikely.” The US spends about 0.4 percent of GDP on defense R&D. China, whose GDP is about a tenth that of the US, would have to spend 4 percent of it on military R&D for many years, and even then would catch up only if the US rested on its laurels.

(Presumably that does not mean that another arms race cannot start — think, for example, about the possibilities of missile defense — but it would take a great deal of warming up. China’s total military expenditure today is barely 2 percent of GDP.)

Next Trajtenberg tallies the composition of current US spending on defense R&D. Big weapons systems command around 30 percent of the total R&D budget — fighter aircraft, intercontinental ballistic missiles, aircraft carriers nuclear submarines and the like. Missile defense research claims around 15 percent. Intelligence gets less than 10 percent; anti-terrorism barely 3 percent. Two “miscellaneous” categories — 33 percent and 9 percent — make up the rest.

All this is prelude to the really interesting part — the analysis of what makes terrorism different from the strategy of “mutually assured destruction.” For one thing, terrorists cannot effectively be deterred by nuclear weapons. Neither can they be destroyed by the use of those weapons.

Terrorists ready from the start to die in the course of their mission are especially difficult to deflect. And the uncertainty of when and where those acts may occur exacts a heavy toll, in terms of both economic and psychological costs.

Here’s where the new meaning of S&L comes in. Trajtenberg distinguishers between measures aimed at preventing terrorism at its source (S), and the steps that may be taken to protect the many locations (L) where terrorism may seek a target — airports, skyscrapers, government buildings, nuclear power plants, server networks and so on.

The centerpiece of Trajtenberg’s approach is a formal model in which terrorism is a three-way game. There are terrorist, who utility depends upon the damage inflicted on their victims. There are their potential targets who, in the absence of terrorism, will go about their business normally, and who therefore are willing to pay something to diminish their risks. And there is the government, which seeks to diminish the risk overall.

It all plays out along a decision tree in which terrorists must decide whether to strike or not and who to hit; in which potential targets can reduce their risks by spending more on their own security; and in which the government must decide how much to spend fighting terrorism at its source, and how much to invest in protecting each potential target.

The key insight here is that two quite different processes are going on. The L strategy — investment at the local level — is essentially a zero-sum game. A dollar spent making one spot safer — makes it more likely that another spot will be attacked. It’s a little like moving into a gated community, or turning yourself into a super-safe airline.

True, Trajtenberg says, each dollar spent on metal detectors, sniffing machines, sky marshals and so on also a delivers a slight deterrent overall. But in his model the overall effect of such jockeying for position is negative. Everyone loses if individuals start investing heavily in their own protection, carrying guns, driving bullet-proof SUVs.

What this means is that defense no longer can be thought of as a global public good. The L-strategy has a more-or-less private component as well — it is a local public good. This has fateful consequences for the cost-effectiveness of various S&L strategies.

It makes sense to spend far more money combating terrorism at its source. In fact, says Trajtenberg, if there are N potential targets, it turns out that S is N times as effective as L. Not surprising, he says, but sobering. Suppose there are 10,000 potentially vulnerable sites?

Thus government should aggressively pursue the S strategy, paying for basic R&D the way it always has. Intelligence-gathering is the key activity here — keeping tabs on would-be terrorists, searching out and disrupting their sources of funding, denying them the opportunity to move freely around the world.

The L-strategy has different implications for funding. National security here has become a partly private good. There’s nothing unprecedented about this, of course; banks have hired their own guards for centuries. But the fact that such private protection can easily enough turn into an arms race of its own is another reason for government to build on what Trajtenberg calls the “cornerstone” of the war on terrorism — the gathering of information on terrorists.

Moreover, there’s plenty of money already in the defense budget to pay for this greatly increased intelligence capability, according to the economist. It just needs to be reallocated from other uses.

The CIA and the code-breaking, eaves-dropping National Security Agency need more resources. Big-ticket items. for which there is no longer any conceivable opponent, will be needed less. A prime example, says Trajtenberg, is the F-22 jet fighter, designed for dog-fights with a generation of jets that the former Soviet Union never built. Gradual upgrades of existing systems should suffice for many decades

It won’t be an easy matter. Existing weapons systems have powerful constituencies of their own. A prime example is the controversy that has broken out about whether to aid Boeing Corp. in its competition with the European consortium that makes Airbus jets, by leasing some of Boeing’s excess production to the Air Force for use as refueling tankers.

All the more reason, however, to keep an eye on the action behind the scenes. Trajtenberg surmises that the different nature of the new R&D may have profound implications for industrial organization. Cold War development led to concentration of R&D and procurement in the hands of a few large corporations, most of them enjoying considerable bargaining power.

In contrast, the new defense technologies that today are on the drawing boards — sensory computer interfaces, computer technologies for massive data analysis, Internet security, public health and biological protection — will be developed on “an entirely different playfield.” Private markets already exist for many of the products sought. A large number of companies have begun to make them.

Thus Trajtenberg concludes that the War on Terrorism may produce a very different style of military-industrial complex than did the Cold War. “R&D programs designed to preserve this diversity and to encourage further competition may prove highly beneficial,” he says — “both for the required defense R&D and for the advanced sectors of the economy themselves, this fostering economic growth.”

And so the large-scale surveillance techniques and intimate man-machine interfaces associated with S-strategy themselves may come to be seen in due course as a “general purpose technology” and an “engine of growth.” Whatever undesirable effects accompany them must remain a topic for another day.