The issue raised by the paper is as follows. "Home equity bias" is a well-known phenomenon in the capital finance literature. That is, despite the existence of global capital markets, it appears that investors around the world disproportionately own home-country equity, rather than globally diversified portfolios. At least in the simple case where, say, the only way to invest in the U.S. is via a "U.S. company," this appears to fly in the face of basic principles of optimal diversification, causing home equity bias to be a "puzzle" badly in need of explanation, and possibly also a problem to be addressed.
The paper's focus diverges from that of typical entries in the home equity bias literature in two respects. First, in an empirical examination of what we know about U.S. home equity bias, it focuses on the question of who (as between Americans and foreigners) ultimately owns the stock of big U.S. multinationals in particular. Second, the underlying reason for being interested does not pertain to explaining or trying to address the puzzle if any. Rather, the paper notes that various arguments that commonly are made in U.S. international tax policy debate - in particular, as it happens, by people on the pro-multinationals side - take it as given that home equity bias is an important fact concerning the ownership of stock in U.S. multinationals.
Suppose, for example, that we are looking at General Electric versus Siemens. These are in some ways similar companies, but GE is legally an American company under our place-of-incorporation rule, while Siemens is legally a German company under their headquarters rule. Perception perfectly matches legal status, however. Most people would say that GE is an "American company" whereas Siemens is a "German company." But in the absence of home equity bias, U.S. individuals (and likewise German individuals) might hold roughly the same ownership percentage in GE as in Siemens.
Indeed, in this non-home equity bias scenario, U.S. individuals might even hold a lower percentage in GE than in Siemens, if holding the latter but not the former helped them to diversify against risks of the U.S. economy that they already bear by reason of working, living, consuming, etc. in the U.S. But then again, since both companies are multinationals with investments around the world, we might well be pushed back, in the absence of home equity bias, towards expecting comparable ownership percentages for U.S. individuals.
It is widely assumed, however, that surely home equity bias does hold here, with the consequence that U.S. individuals are assumed to hold a MUCH higher percentage of GE than Siemens. Just to capture what I am guessing might be the standard guess, perhaps the common view would hold that GE is, say 80% owned by American individuals, and Siemens, say, less than 5%. But the paper says: "Not so fast - how confident should we actually be about this empirical claim?"
While the paper's main point concerns the lack of data that is actually reliable (despite multiple data sources), let's start by looking behind the data for a moment. Suppose that, for all U.S. corporate equity, as defined by state incorporation statutes and including everything that is closely held, you could figure out the percentage that is owned by U.S. individuals. Reflecting that some people who live in the U.S. incorporate in order to own and operate their own businesses, one would come up with evidence that appeared to show "home equity bias" but that actually conflated that phenomenon with the distinct one of what people do with their investment portfolios. Now, just as a silly thought experiment, suppose we passed a law requiring all restaurants, fast food places, billiards parlors, and dry cleaning establishments to incorporate. "Home equity bias" would seem to have increased, whereas in fact nothing of substance would actually have changed at the margin that interests us here.
It's one thing, already, to say that we're interested in publicly traded companies (which at least some of the data sources that the paper discusses generally don't break out as a separate category). But the paper notes that multinational companies in particular are the poster child for those as to which you might expect LESS home equity bias, under most of the more credible explanations of the phenomenon in the literature. So the paper argues, after going through the soft spots in one data source after another, that we really don't have any convincing evidence that stock in prominent U.S. multinationals is disproportionately owned by U.S. individuals.
A small additional word in furtherance of the paper's skeptical view of what we actually know about this: Suppose we actually knew who literally owns all the stock - including indirectly, by reason of looking through mutual funds and all such other intermediate entities to the ultimate individual level. Even so, in a world full of derivative financial instruments (options, swap contracts, forward contracts, repo agreements, etc.), this might fall well short of telling us who really bears the economics associated with ownership of a particular company's stock. For example, if you own the stock but we use a swap to transfer most of the economics of the share's performance to me, the data about legal ownership will be getting it wrong. Derivatives, side bets, and all the rest make it even harder to figure out the underlying fundamentals than the paper's detailed and skeptical critique of the various leading data sources already suggests.
OK, why might all this matter? The main reason I think that it matters who "owns" (in the relevant economic sense) the stock of U.S. as compared to foreign multinationals, from the standpoint of U.S. international tax policy debate, is that (a) the entity-level corporate tax is a mechanism for indirectly imposing income taxation on the ultimate owners, and (b) we both do and should think very differently about taxing (whether directly or indirectly) resident individuals as opposed to foreign individuals.
For resident individuals, the main aim is to allocate tax burdens on the basis of some measure of some measure (e.g., income) that at least proxies for some underlying attribute such as ability or material wellbeing. Hence, if we could, we would like to include, say, all of Bill Gates' income in the tax base, whether he earns it at home or abroad, and/or through a corporation or directly.
For foreign individuals, the idea is that it would be nice from, a domestic national welfare standpoint, to get $$ from them if we can. So if we impose taxes on their activity in the US or through US corporations, and they bear this tax as a matter of economic incidence, then we are ahead of the game unless the associated cost to U.S. individuals (deadweight loss, decline in positive externalities from transactions with foreigners, etc.) is disproportionately high. But this is a very different proposition from that for resident individuals, and in practice it may call for being resolved differently. (For example, we probably don't want to tax the foreign individuals at all, in settings where we, not they, bear the economic incidence of the tax.)
Anyway, that's why I think the ownership data would be of interest, as one more potential input into evaluating all the issues in international tax policy. (It's also of interest, of course, as a puzzle or non-puzzle in the capital finance literature.) The main angle pursued in the Sanchirico paper, however, is somewhat different - not reflecting any particular disagreement, but merely because he is hunting different prey.
As the paper notes, pro-U.S. multinational proponents sometimes argue that the presumed predominantly American share ownership of U.S. multinationals provides support for (a) lessening the companies' U.S. tax burdens on competitiveness grounds and (b) viewing foreign dividend repatriation tax holidays as likely to increase U.S. investment. The argument on (a) is that U.S. individuals are the shareholders who putatively are getting enriched. The argument on (b) might go as follows: Suppose a foreign repatriation tax holiday is expressly permitted to fund domestic dividends and share repurchases. Shareholders who were U.S. individuals ostensibly would be more likely to plow the money received back into the U.S. economy.
The paper says, in effect: Let's accept these contentions arguendo and ask if the underlying factual premise is true: Are the U.S. multinationals' shareholders disproportionately American. We really don't know this, and people therefore shouldn't just assume it, thereby weakening the underlying arguments.
I myself, while recognizing that these arguments are being made in the U.S. international tax policy debate, don't regard them as strong ones - no matter how one comes out overall on the underlying issues. Thus, I would be skeptical of their value and weight even if I were to accept (as I previously had been inclined to) the asserted fact of predominantly American share ownership of U.S. multinationals.
For example, I note that the "competitiveness" argument is not easily translated into a compelling claim that people who own U.S. multinationals' stock will earn extra-normal returns if, but only if, the U.S. doesn't tax the companies' foreign source income. And I consider tax holidays a horrible idea because they give companies and investors the lesson that you should just wait for the next holiday. Moreover, I would not expect the holidays to boost the U.S. economy unless we were to accept a not very compelling story in which liquidity constraints (at the shareholder level and/or the domestic firm level) are the big problem that is holding back national macroeconomic performance.
Overall, however, the paper offers a nice lesson in the merits of learning that perhaps you know less than you thought you knew.
It's one thing, already, to say that we're interested in publicly traded companies (which at least some of the data sources that the paper discusses generally don't break out as a separate category). But the paper notes that multinational companies in particular are the poster child for those as to which you might expect LESS home equity bias, under most of the more credible explanations of the phenomenon in the literature. So the paper argues, after going through the soft spots in one data source after another, that we really don't have any convincing evidence that stock in prominent U.S. multinationals is disproportionately owned by U.S. individuals.
A small additional word in furtherance of the paper's skeptical view of what we actually know about this: Suppose we actually knew who literally owns all the stock - including indirectly, by reason of looking through mutual funds and all such other intermediate entities to the ultimate individual level. Even so, in a world full of derivative financial instruments (options, swap contracts, forward contracts, repo agreements, etc.), this might fall well short of telling us who really bears the economics associated with ownership of a particular company's stock. For example, if you own the stock but we use a swap to transfer most of the economics of the share's performance to me, the data about legal ownership will be getting it wrong. Derivatives, side bets, and all the rest make it even harder to figure out the underlying fundamentals than the paper's detailed and skeptical critique of the various leading data sources already suggests.
OK, why might all this matter? The main reason I think that it matters who "owns" (in the relevant economic sense) the stock of U.S. as compared to foreign multinationals, from the standpoint of U.S. international tax policy debate, is that (a) the entity-level corporate tax is a mechanism for indirectly imposing income taxation on the ultimate owners, and (b) we both do and should think very differently about taxing (whether directly or indirectly) resident individuals as opposed to foreign individuals.
For resident individuals, the main aim is to allocate tax burdens on the basis of some measure of some measure (e.g., income) that at least proxies for some underlying attribute such as ability or material wellbeing. Hence, if we could, we would like to include, say, all of Bill Gates' income in the tax base, whether he earns it at home or abroad, and/or through a corporation or directly.
For foreign individuals, the idea is that it would be nice from, a domestic national welfare standpoint, to get $$ from them if we can. So if we impose taxes on their activity in the US or through US corporations, and they bear this tax as a matter of economic incidence, then we are ahead of the game unless the associated cost to U.S. individuals (deadweight loss, decline in positive externalities from transactions with foreigners, etc.) is disproportionately high. But this is a very different proposition from that for resident individuals, and in practice it may call for being resolved differently. (For example, we probably don't want to tax the foreign individuals at all, in settings where we, not they, bear the economic incidence of the tax.)
Anyway, that's why I think the ownership data would be of interest, as one more potential input into evaluating all the issues in international tax policy. (It's also of interest, of course, as a puzzle or non-puzzle in the capital finance literature.) The main angle pursued in the Sanchirico paper, however, is somewhat different - not reflecting any particular disagreement, but merely because he is hunting different prey.
As the paper notes, pro-U.S. multinational proponents sometimes argue that the presumed predominantly American share ownership of U.S. multinationals provides support for (a) lessening the companies' U.S. tax burdens on competitiveness grounds and (b) viewing foreign dividend repatriation tax holidays as likely to increase U.S. investment. The argument on (a) is that U.S. individuals are the shareholders who putatively are getting enriched. The argument on (b) might go as follows: Suppose a foreign repatriation tax holiday is expressly permitted to fund domestic dividends and share repurchases. Shareholders who were U.S. individuals ostensibly would be more likely to plow the money received back into the U.S. economy.
The paper says, in effect: Let's accept these contentions arguendo and ask if the underlying factual premise is true: Are the U.S. multinationals' shareholders disproportionately American. We really don't know this, and people therefore shouldn't just assume it, thereby weakening the underlying arguments.
I myself, while recognizing that these arguments are being made in the U.S. international tax policy debate, don't regard them as strong ones - no matter how one comes out overall on the underlying issues. Thus, I would be skeptical of their value and weight even if I were to accept (as I previously had been inclined to) the asserted fact of predominantly American share ownership of U.S. multinationals.
For example, I note that the "competitiveness" argument is not easily translated into a compelling claim that people who own U.S. multinationals' stock will earn extra-normal returns if, but only if, the U.S. doesn't tax the companies' foreign source income. And I consider tax holidays a horrible idea because they give companies and investors the lesson that you should just wait for the next holiday. Moreover, I would not expect the holidays to boost the U.S. economy unless we were to accept a not very compelling story in which liquidity constraints (at the shareholder level and/or the domestic firm level) are the big problem that is holding back national macroeconomic performance.
Overall, however, the paper offers a nice lesson in the merits of learning that perhaps you know less than you thought you knew.
No comments:
Post a Comment