Economics has its founding fathers, like Adam Smith and David Ricardo. If list of greats were compiled for finance, Harry Markowitz would number among them. Markowitz helped invent Model Portfolio Theory, a mathematically elegant approach to optimizing investment portfolios that considers not only how well one expects investments to do, but also how certain one is of ones judgments, and the typical inter-relationships between "surprises" in different investments.

Markowitz's theory of investment and Ricardo's theory of trade are meant for very different domains, but the contrast between them is striking. Ricardo teaches that nations should determine their comparative advantage, and specialize in that, trading for what they do not produce most efficiently. Markowitz suggests that investors beware the temptation of specialization, and diversify even into apparently inferior investments to limit risk. Both are revered. What explains the difference?

Nations, after all, are investors, in the aggregate. Specializing in manufacturing or software or wine or cloth entail vastly different portfolios of fixed structures and human training. And investors, like nations, have comparative advantages. Consider a venture capitalist, or an "angel" investor. Each faces different investment opportunity sets, determined by who they know and the localities or domains in which they invest. Every investment has an opportunity cost (the expected return on other investments). Using return estimates alone, most investors would find that choosing one or a very few investments would uniquely maximize expected return, net the cost of foregone opportunities. The situation of a nation that can trade for what it does not produce is broadly similar to that of an investor, for whom a high return in a software venture yields specie that can be used to purchase food, shelter, and yachts. (The no-trade analogy would be an investor whose profits at a stocking factory would have to be taken in the form of inedible socks.)

So, should nations and investors both specialize? Or both diversify? Or should they behave differently, and if so why? From an investment perspective, the question would turn on two considerations, uncertainty and the optimal locus for diversification. Uncertainty is the easiest factor: An investor who can perfectly predict the real returns on investments should absolutely not diversify, even under Markowitz's theory. Ricardo took the comparative advantages — that is the relative investment returns — of nations as fixed and given. Under these circumstances, Markowitz would agree that specialization is the way to go. But how good are nations, really, at knowing their comparative advantages? And even when they choose correctly, how likely is it that changes in circumstance or random surprises render judgements inaccurate? The greater the uncertainty, either of estimation or environment, the greater the case that nations should diversify.

A second question hinges on who is best placed to diversify. In the 1960s, it became fashionable among corporations to form "conglomerates", large groups of unrelated businesses held under a single corporate umbrella. There were a bunch of rationales for this, including the ability of subsidiaries to raise capital internally, economies of scale in shared functions, and the market and political power associated with size. But one important motive for conglomeration was Markowitz' portfolio theory. It was argued that a diversified firm faced fewer risks than a specialized one, since when one industry was fairing poorly, other business units might do well to make up the difference.

Conglomerates are no longer fashionable. Investors rejected the case for corporate diversification, because they were perfectly capable of diversifying themselves. Why should MegaCorp include unrelated businesses A, B, and C, when an investor can divide her own portfolio between stock in A, B, and C? In fact, MegaCorp's diversification harms the investor, because the investor loses the opportunity to customize exposure to A, B, and C according to her own circumstance and expectations.

Unfortunately, most citizens can't hedge their exposure to nations in the way that investors can diversify investments in firms. This is a strong argument in favor of diversifying at the national level, even when it cuts against maximizing comparative advantage.

There are a several ways that nations may be different than investors that buttress the case for national specialization. Investors choose their investments according to flawed analytical processes, and make mistakes. When a capitalist economy specializes under open trade, it is a vast market that decides in which fields to specialize, what factories to build, what competences to educate. Perhaps owing to "the invisible hand", "the wisdom of crowds", or whatever, market economies make such better choices than private investors that there effectively is no uncertainty. The market always chooses correctly, so specialization is the optimal choice. I find this argument unpersuasive, both because I think markets can be short-sighted and mistaken, and because I believe in irreducible uncertainty that no predictive institution can overcome. Another difference between nations and ordinary investments is that national investment has dynamic effects. When a nation begins to specialize in, say, electronics manufacturing, economies of scale and network effects kick in that serve to magnify any original advantage the nation had in that field, and turn the investment into a kind of self-fulfilling prophecy. This, I think, is a quite reasonable point. But the difference between investors and nations can be drawn too starkly. A venture capitalist or angel investor also works, post-investment, to advise firms, to recruit good people, and to create connections between complementary firms. But these kinds of investors still do diversify.

If one buys the case that nations are like investors, and that investors ought diversify in an uncertain world, what does that mean in practical terms? It depends. It might mean nothing. Perhaps the ordinary functioning of national markets produce sufficiently diversified national portfolios on their own, in which case all is well and good. But it does suggest the possibility of market failure. If national investment is skewed or concentrated in some fashion whose future performance (relative to other possible portfolios) is uncertain, public action to promote diversification might be reasonable.

"Industrial policy" is unfashionable, and it is of course true that political processes are flawed and corruptible. The case I've made opens the door for rent-seeking operators to seek government handouts in the form of "diversification subsidies". Any policy based on this argument would have to be designed with great care. Nevertheless, that chemotherapy is poisonous does not imply that cancer does not exist.

Note: While David Ricardo long ago joined Zeus on Mount Olympus, Harry Markowitz, whose work I much admire, is a living, active scholar. The views expressed in this essay are entirely my own, and if I seem to have put words in Harry Markowitz's (or David Ricardo's) mouth, I do apologize.