Globalization and Capital Markets

A resurgence of hostility to financial integration, to which it offers a quantitative long-run record plus an organising framework (§1). The opening is dated precisely: “At the turn of the twenty-first century, the merits of international financial integration are under more forceful attack than at any time since the 1940s,” citing Bhagwati on the risks outweighing the benefits and Eatwell’s claim that free capital flows have been “associated with a deterioration in economic efficiency (as measured by growth and unemployment).” The authors treat the backlash as understandable given the crises of 1994-95, 1997-98 and 2001-02, and explain why international finance is distinctively fragile: transactions “rely intrinsically on the expectation that counterparties will fulfill future contractual commitments,” while “oversight, adjudication, and enforcement all are orders of magnitude more difficult among sovereign nations with distinct national currencies than within a single national jurisdiction,” and “there is no natural world lender of last resort.” Their stated contribution is twofold: to document the vicissitudes quantitatively, and to argue that “theories of how international capital mobility has evolved must be understood within the framework of the basic policy trilemma constraining an open economy’s choice of monetary regime.”

A constraint on feasible policy combinations, offered as a proximate rather than a deep explanation (§1.3). “An open capital market deprives a country’s government of the ability simultaneously to target its exchange rate and to use monetary policy in pursuit of other economic objectives,” so at most two of the three goals – free capital movement, a fixed rate, an independent monetary policy – can be had. The historical proposition follows: mobility “has prevailed and expanded under circumstances of widespread political support either for an exchange-rate-subordinated monetary regime (for example, the gold standard), or for a monetary regime geared mainly toward domestic objectives at the expense of exchange-rate stability (for example, the recent float),” while the middle ground “has, almost as a logical consequence, entailed exchange controls or other harsh constraints.” Two qualifications are made explicitly. The fixed-versus-floating choice is not binary: “the degree of exchange-rate flexibility lies on a continuum … [and] the notion of a ‘free’ float is an abstraction with little empirical content, as few governments are willing to set monetary policy without some consideration of its exchange-rate effects.” And the trilemma does not identify the driver: among Cohen’s four categories – technology, state-interest policy competition, domestic politics, and ideology – the authors regard technology as secondary for their period and “explanations along the lines of (ii) and especially (iii) as the ‘deep factors’,” with the trilemma serving only to “constrain the choice set within which the deep factors play their roles.”

Because every candidate criterion is only valid jointly with auxiliary assumptions, which the authors demonstrate with two counterexamples (§2). Price convergence fails because “prices would be identical in two identical neighboring economies … but if the two markets were physically separated by an infinitely high transaction-cost barrier one could hardly describe them as being integrated in a single market, as the equality of prices was merely a chance event.” Flow volumes fail symmetrically: remove the barrier between those same economies and “we would then truly have a single integrated market, but, since on either side of the barrier prices were identical in autarky, there would be no incentive for any good or factor to move.” Hence “convergence of prices and movements of goods are not unambiguous indicators of market integration,” and the strategy is to use “a battery of tests, using both quantity and price criteria of various kinds,” with the argument resting on agreement across them. The sample constraint is stated openly: “we will frequently be restricted to looking at between a dozen and twenty countries for which long-run macroeconomic statistics are available, and this sample will be dominated by today’s developed countries.”

As foreign capital over nominal GDP, a second-best normalisation whose weaknesses the authors set out before using it (§2.1). The ideal denominator would be the total capital stock, but financial-capital measures “have greatly multiplied over the long run as financial development has expanded the number of balance sheets,” and real capital stock data are available for few countries and usually in constant rather than current prices. Output is used instead, on the assumption that the capital-output ratio has not changed much – “we have little firm evidence to suggest that it has,” with the conventional wisdom putting it at 3 to 4. The GDP series come from Maddison’s constant-price 1990 dollar estimates reflated by a US deflator, an approach the authors call “crude, since, in particular, it relies on a PPP assumption.” The numerator is the harder problem: the IMF only began reporting international investment positions in 1997, “beginning in 1980 for less than a dozen countries, and expanding to about 30 countries by the mid-1990s,” and stock data “are not simply a temporal aggregate of flows” but depend on “past flows, capital gains and losses, any retirements of principal or buybacks of equity, defaults and reschedulings.” The resulting series: foreign assets “just 7 percent of World GDP” in 1870, “just below 20 percent in the years 1900-14,” then 8 percent in 1930, 11 percent in 1938, 5 percent in 1945, 6 percent in 1960, 25 percent in 1980 and 62 percent in 1995.

Because the two bracket the truth from opposite directions, and the gap between them is where the interesting result lies (§2.1). The world-GDP ratio is a lower bound, since countries with no foreign-investment data contribute zero to the numerator but their output to the denominator. The sample-GDP ratio – restricting the denominator to countries with numerator data – is an upper bound, because historical collection “has usually focused on the principal players.” The two are “very close after 1980” but “before 1945 they are quite far apart: from 1870 to 1914, the sample of seven countries has a foreign asset to GDP ratio of over 50 percent, far above the ‘world’ figure of 7 to 20 percent. By this measure we only surpassed the 1914 ratio as recently as 1990, and only narrowly even then.” The interpretation the authors draw concerns portfolio diversification rather than mobility per se: “these seven major creditors were exceptionally internationally diversified in the late nineteenth century in a way that no group of countries is today,” so “in countries like today’s United States, we still have yet to see a return to the extremely high degree of international portfolio diversification seen in, say, Britain in the 1900-14 period.”

A single dominant creditor before 1914, far more concentrated than anything since (§2.1). “For all of the nineteenth century, and until the interwar period, the British were rightly termed the ‘bankers to the world’; at its peak, the British share of total global foreign investment was almost 80 percent. This is far above the recent U.S. share of global foreign assets, a mere 22 percent in 1995, and still higher than the maximum U.S. share of 50 percent circa 1960.” The only earlier rival was the Dutch, holding “perhaps 30 percent of global assets in 1825.” The US was a debtor rather than a creditor in that era, becoming a large creditor abruptly through European wartime borrowing, though “the dislocations of the interwar years were to postpone the United States’ rise as a foreign creditor.” A footnote is careful to distinguish gross from net: “this is the gross foreign investment position, not the net position. The United States is also now the world’s number one debtor nation, in both gross and net terms, having become a net debtor for the first time since the First World War in the late 1980s.”

That the mid-century collapse must reflect impediments to flows rather than changes in the desirability of holding foreign assets – stated as a conditional (§2.1, §2.5). They acknowledge that “[f]iguring whether too much or too little diversification existed at any point must remain conjectural, and conclusions would hinge on a calibrated and estimated portfolio model applied historically.” The argument is therefore by elimination: “unless the global economy has dramatically changed in terms of the risk-return profile of assets and their global distribution, we have no prior reason to expect the efficient degree of diversification to have changed,” so “unless a massive such change did occur in the 1914-45 period, and unless it was then promptly reversed in the 1945-90 period, we cannot explain the time path of foreign capital stocks … except as a result of a dramatic decline in capital mobility in the interwar period, and a very slow recovery thereafter.” The §2.5 summary states the general version, again with a concession: changes in the degree of mobility took the form of “changes in the impediments to capital flows, rather than any encouragement or discouragement to flows arising from structural shifts within the economies themselves. That is not to discount the fact that such changes have occurred, and are no doubt important at the margin.”

Stationary differentials over more than a century, contradicting much of the earlier literature – on ex post real rates with a stated and acknowledged proxy problem (§2.2). Nominal rates are monthly long-term government bond yields of seven years or longer from Global Financial Data; inflation is the ex post 12-month forward change in the CPI; the real rate is the difference, “and for now we make the standard assumption that this is equal to the ex ante real rate plus a white-noise stationary forecast error.” Long rates are chosen because they are “most directly related to financing costs for capital investments,” and because slow real exchange rate mean reversion makes expected exchange rate changes hard to discern at short horizons. The authors flag the compromise: “in measuring long-term real interest rates, we would like to proxy long-term inflation expectations but that cannot be done reliably. Thus we follow earlier empirical studies in utilizing a relatively short-horizon inflation measure notwithstanding the longer term of the corresponding nominal interest rates.” With war years and the German hyperinflation excluded, “differentials have varied widely over time, but have stayed relatively close to a zero mean,” and formal ADF and DF-GLSu tests reject the unit-root null “in almost all cases at the 1 percent level in all periods except for the recent float,” where the evidence is stronger for 1986-2000 than 1974-86. The authors note this “contradicts much of the empirical literature produced through the mid-1990s” and attribute the difference to earlier work using only the recent float and shorter samples. On volatility across regimes, the finding that there is “perhaps very little change between the pre-1974 period and the float” is reconciled with Baxter and Stockman’s result that, exchange rates aside, “there is little difference in the behavior of macro fundamentals between fixed and floating rate regimes.”

A long-run series spliced from two different arbitrage operations, with an explicit correction for period-specific transaction taxes (§2.3). Since onshore-offshore comparison is impossible for much of the period, pre-1920 integration is measured using the long bill of exchange: the implied offshore sterling rate in New York, derived from the spot rate and the dollar price of sterling deliverable in 60 days, is compared to the London market discount rate. The series shows differentials “relatively small and steady under the pre-1914 gold standard,” opening with the First World War, staying large in the early 1920s, narrowing briefly late in that decade, widening sharply in the early 1930s, large again from the late 1940s through the mid-1950s (with a Suez spike), shrinking, reopening around sterling’s 1967 devaluation, and small again since UK controls went in 1980 – “even smaller now than before 1914.” The systematic pre-1914 excess return in New York is not read as segmentation but decomposed: purchasers of a long bill factored in “commissions, profit margins, and, importantly, the stamp duty (0.05 percent of the bill’s face value),” which for a 60-day bill amounts to about 75 basis points annualised, and “[b]y subtracting that ’tax’ from the pre-1914 differentials … the apparent average excess return in New York disappears” – indeed becomes negative for 1890-1914, “so that 75 basis points in additional costs may well be an overestimate.” A footnote lists five further data problems (different arbitrage operations before and after 1920, varying maturities, mixed end-of-month and weekly averages, administered wartime prices, imperfect time alignment) and states the modest aim: “to convey a broad sense of the trend in integration, not to pursue a detailed hunt for small arbitrage possibilities.”

The width of the estimated no-arbitrage bands (§2.3). Keynes and Einzig held that in the interwar period “at least a 50 basis point covered differential would be needed to induce arbitrage,” and Peel and Taylor’s threshold-autoregressive work on weekly data confirms a band close to that. The authors’ own first pass gives “a band of inaction of plus or minus 60 basis points” for dollar-sterling between June 1925 and June 1931 and “plus or minus 91 basis points” for interwar sterling-mark, against “bands of only plus or minus 19 basis points for New York-London and plus or minus 35 basis points for London-Berlin” on 1880-1914 differentials, after removing a constant mean. For contemporary comparison, Clinton puts the mid-1980s threshold at about 6 basis points and Balke and Wohar 50 percent higher for 1974-93. The authors suspect their pre-1914 bands are still too wide, and conclude that “the degree of integration among core money markets achieved under the classical gold standard must be judged as truly impressive compared to conditions over the following half-century or more,” with the Great Depression standing out “as an event that transformed the world capital market.”

It agrees for the G7, on returns data the authors are careful to flag as afflicted by survivorship bias and missing dividends (§2.4). The sample is up to 22 country stock price indices in annual dollar returns since 1880, but “the sample size diminishes markedly prior to 1950, evidence of the survivorship problem” – about a dozen countries in the interwar period, five to ten before 1920, and only three G7 series before 1920 – and “[s]ince dividend data are not available for the entire sample, the calculated returns are based on equity-price changes only.” With those caveats, the ten-year rolling cross-sectional standard deviation of G7 returns is “strikingly consistent with the U-shape hypothesis”: “[r]eturns showed relatively little dispersion prior to 1914, but larger gaps opened up in the interwar period. This dispersion reached a peak around 1945 or 1950, but has been falling since, with a minor reversal in the late 1960s, but convergence again after 1980.” The authors also situate their exercise against Jorion and Goetzmann’s finding that most markets returned around 1 percent in real terms over the century “with the exception of the United States, which has yielded around 4 percent annually since 1921” – figures which “caution that U.S. exceptionalism might extend to stock market returns also.”

Because the timing is wrong and the direction is monotone (§2.5). “Clearly, technology is a poor candidate for the explanation of the twentieth century collapse of capital mobility. In the 1920s and 1930s, the prevailing financial technologies were not suddenly forgotten by market participants: indeed some technologies, such as futures markets for foreign exchange came to fruition in those decades. Technological evolution was not smooth and linear, but, as we have already noted, was at least unidirectional, and, absent any other impediments, would have implied an uninterrupted progress toward an ever more tightly connected global marketplace.” What is left is policy: “the shifting forces of national economic policies, as influenced by the prevailing economic theories of the day, loomed large during and after the watershed event of the twentieth century, the economic and political crisis of the Great Depression.”

A consistent London-market panel for 1870-1940 that finds the interwar gold standard much less credible than its predecessor, in contrast to Bordo, Edelstein and Rockoff (§3.1). The earlier studies could not be merged: Bordo-Rockoff used secondary-market yields to maturity on long-term government bonds in London, while Bordo-Edelstein-Rockoff used new issues in New York and their flotation yields, with “a small sample that was often interrupted by missing data in years when no issues took place … one that could raise a potential sample-selection issue.” The authors instead take Global Financial Data yields on London-traded government bonds “denominated in gold or in sterling” – to isolate default from currency risk – for 20 countries split into core-and-empire and periphery. For 1870-1913, in the preferred AR(1) specification, “[b]eing on gold appears statistically significant before 1914, contributing about minus 57 basis points to the spread over London,” and “it is only the peripheral countries whose public debt levels appeared to increase spreads,” with “a 10 percent rise in a peripheral country’s ratio of debt to GDP add[ing] 7.2 basis points to its borrowing cost” – sizeable against a consol yield of about 3 percent. Inflation is “qualitatively small and of low statistical significance” prewar, and the autoregressive parameter of 0.68 suggests “countries could rely on some reputational persistence.” For 1925-30 the picture changes: gold at prewar parity is “statistically insignificant, or even incorrectly signed”; there is “no interwar evidence that the markets differentiated between core and periphery countries in evaluating the country-risk implications of rising debt burdens,” with a 10-point debt rise raising spreads by up to 43 basis points (OLS) or 22 (AR(1)), so “the sensitivity of bond spreads to debt was about five times larger after the war than before”; and the autoregressive parameter falls to 0.30. One result the authors call “unusual” and do not oversell: “returning to gold at a devalued parity is estimated to have a more beneficial effect than returning at prewar parity,” consistent with Drazen and Masson’s view that draconian policies can damage credibility – “[t]he results are not conclusive, but further research is perhaps warranted.” They also note an unreconciled tension with Hallwood, MacDonald and Marsh’s evidence of a credible late-1920s gold standard, conjecturing that “the bond markets adopted a longer perspective.”

Enfranchisement and macroeconomic activism closing off the gold-standard corner, then trade success reopening the trilemma in the 1970s (§3.1, §3.2). For the collapse, “the major political economy forces at work during this period were increasing pressure for macroeconomic activism, particularly from newly- or better-enfranchised groups such as the working classes,” with the consequence that “[i]f fixed exchange rates were to be maintained, then capital mobility would have to be compromised.” For the recovery, the mechanism is that trade integration undermined capital controls: postwar policymakers “successfully promoted growing world trade,” and “[b]y the late 1960s, the very success of these initiatives in forging trading linkages among countries simultaneously made capital flows across borders ever more difficult to contain. As a result, the trilemma re-emerged with full force, and on a global scale, in the early 1970s.” Floating rates then “helped reconcile the social demand for domestic macroeconomic stabilization with the interest of the business community for open markets in goods and assets,” while Europe’s single-market and single-currency project and other regions’ hard pegs and dollarization represent the other corner, “in either case bending to the trilemma by giving up monetary policy autonomy.” On the developing world the authors are candid about US pressure and its consequences: “[i]n part a reflection of U.S. business interests, American administrations have pushed developing economies to liberalize on capital account; in some cases, liberalization ran far ahead of domestic financial systems’ absorptive capacities, and clashed with national exchange rate policies. The resulting contradictions helped spark the developing country currency crises of the latter 1990s.”

That today’s large gross positions net out to very little, so integration finances diversification rather than development (§3.2). Pre-1914 flows were “long-term investment capital, and virtually unidirectional at that,” the United States being the notable exception, so creditors “developed enormous one-way positions in their portfolios” and “[t]o a first approximation, the gross asset and liability positions were very close to net in that distant era.” Now “the rank of countries by foreign assets in the IMF data is very highly correlated with the rank by foreign liabilities,” with Britain, Japan, Canada, Germany and the Netherlands all large on both sides, and “since 1980, the net foreign asset position (or liability) positions in the world economy have remained very low indeed” – indeed “the picture is one of relative decline in the size of net foreign capital stocks relative to GDP” if the asset data are trusted more. Hence: “[t]oday’s foreign asset distribution is much more about asset ‘swapping’ by rich countries – diversification – than it is about the accumulation of large one-way positions … It is therefore more about hedging and risk sharing than it is about long-term finance and the mediation of saving supply and investment demand between countries. In the latter sense, we have never come close to recapturing the heady times of the pre-1914 era, when a creditor like Britain could persist for years in satisfying half of its accumulation of assets with foreign capital, or a debtor like Argentina could similarly go on for years generating liabilities of which one half were taken up by foreigners.” The authors connect this directly to Feldstein and Horioka: “still to a very great extent today, a country’s net wealth will depend, for accumulation, on the provision of financing from domestic rather than foreign sources.”

Their share of global liabilities has fallen sharply since 1900, and the authors list competing interpretations without choosing (§3.2). “In 1900, LDCs in Asia, Latin America, and Africa accounted for 33 percent of global liabilities, in the 1990s only 11 percent,” so “[t]he global capital market of the nineteenth century centered on Europe, especially London, extended relatively more credit to LDCs than does today’s global capital market.” Candidate explanations are set out as questions: “did Britain, as an imperial power, favor LDCs within her orbit with finance? or, today, does the global capital market fail in the sense that there are insufficient capital flows to LDCs, and an excess of flows among developed countries? These are hard claims to prove, as market failure could be a cause, as could a host of other factors including institutions and policies affecting the marginal product of capital in different locations.” They also note the accounting point that the result partly follows mechanically from rich countries topping both rankings. On the distribution of recipients, pre-1914 capital “was distributed bimodally; it moved to both rich and poor countries, with relatively little in the middle” – to settler economies for abundant land and to poor countries for abundant labour – whereas today’s creditor-debtor pairs “are more rich-rich than rich-poor,” reproducing “the paradox noted by Lucas (1990), of capital failing to flow to capital-poor countries.” The comparison is if anything understated, since “a century ago world income and productivity levels were far less divergent than they are today, so it is all the more remarkable that so much capital was directed to countries at or below the 20 percent and 40 percent income levels.”

That the benefits of borrowing and lending are real but that the constraints and risks are too, and that poorer countries need more foreign capital than they now receive (§1.5, §3.2). On the benefits: “the ability to lend or borrow represents, trivially, a loosening of constraints relative to a perfectly closed economy. In this dimension, at least, open trade in financial markets offers unambiguous gains relative to a closed economy. Such trades permit insurance, the smoothing of shocks, and allow capital to seek out its highest rewards.” On the costs: mobility “raises concerns, particularly for policymakers attached to certain policy goals that may be inconsistent with the free flow of capital,” and “the risks of financial and balance of payments crises – some of them self-fulfilling crises unrelated to ‘fundamentals’ – may represent further obstacles.” The closing judgement is conditional rather than triumphal: capital is kept out of poorer countries by controls, by risk perceptions formed “after a century of exchange risks, expropriations, and defaults,” and by domestic distortions, but “[p]oorer countries must draw on foreign capital to a greater extent than they do at present if they are to achieve an acceptable growth in living standards. That is a fundamental reason why reform and liberalization in the developing world, despite the setbacks of the late 1990s, are likely to continue, albeit hopefully with due regard to the painful lessons learned in the recent past.”