The second industrial revolution, based upon electricity, chemicals, and the internal combustion engine, and characterized by capital intensity and large initial investments, was made possible by the emergence of corporate securities markets. Markets allowed banks to transform long-term loans to industry into securities, recoup their liquidity, and lend anew. The development of securities markets in general was made possible by the agglomeration of savings in the core. Secondary securities markets needed a lot of liquid assets to function well. Yet, most securities markets did not find the cash they needed to grow and, as a result, stagnated. Securities markets were starved of cash in countries where local credit sectors carved up a sizable portion of the deposit market. Stock markets in the second half of the nineteenth century constituted, along with large commercial banks, a new “corporate finance,” geared to the financial needs of the new industrial sectors. Land and other traditional sectors, in contrast, had no use for “corporate finance,” but, instead, were banking with the non-profit sector (savings banks, credit cooperatives, and mortgage banks). The two financial sectors were in competition for resources, mainly deposits. The competition was adjudicated politically, through regulation. Since the size of the local non-profit sector was a negative function of the degree of centralization of the state, the development of corporate securities markets conversely was a positive function of state centralization.

Securities markets are growing again. From 1980 to 1996, listed stock capitalization doubled in the United States, quadrupled in France, Germany, and the UK, and increased by a multiple of twenty-six in Portugal, though, in this case, the initial value was very low. Although most of this increase was spent recovering from two poor decades – US market capitalization, for instance, did not regain its 1961 GDP-weighted value until 1993 – the surge is sufficiently pronounced and ubiquitous enough to be considered a qualitative change.

Market growth implies a relative decline in the role of banks, with predictable redistributional effects. Disintermediation reduces the twofold capacity that the financial system has to overcome information asymmetry and insure firms against cyclical financial vicissitudes. Although there is no evidence of a “Macmillan gap” yet, the trend toward securitization has the potential to hurt firms that are small or engaged in traditional sectors. We should not be surprised, therefore, to observe that disintermediation and securitization are not pursued to the same length in all countries, but, instead, that the largest stock markets, today as prior to the Great Depression, are found in financial systems that exhibit a low degree of segmentation and in countries with centralized state structures.

Still, we should not expect the 1913 patterns to reemerge intact.

The period from 1960 until the present is characterized by a strengthening of core–periphery tensions in finance. Deregulators are removing the protective measures that were established in the wake of the Great Depression; they are also scaling back the state credit sector, which in many countries reached unprecedented proportions in the immediate postwar period.

However thorough and ubiquitous, the present trend is unlikely to bring about the convergence of OECD banking systems on the money-center bank model; it is of a more limited character. Deregulation has merely restored the degree of competition that existed in pre-Depression days. The center banks have managed to regain a sectoral dominance in countries where they were dominant before World War I – in centralized countries. In contrast, center banks have remained a junior partner in countries where the savings and cooperative banks constituted the dominant sector – in decentralized countries. Certainly, concentration has increased across countries, but concentration has taken place only within sectors, not across them.

The interwar and early postwar decades

Ongoing financial deregulation has consisted in the removal of four decades of regulatory layers. These layers took two forms: (1) the promotion of state credit and (2) the regulation of banking competition.

State credit

State credit is the allocation of credit by the central government through so-called specialized credit banks or state credit banks, which finance their needs by issuing state-guaranteed bonds. State credit banks are different from postal (and other state-run) savings banks.

Modern banks were founded in the nineteenth century. Unlike Gerschenkron and, more generally, the Saint-Simonian approach to credit, which views the origins of these banks as a response to industrialization, I argue that these banks were the outgrowth of an innovation in the payments mechanism. It consisted in the gradual displacement of commercial paper, traded at a discount in a centrally located market, by drafts and deposits, offered by a handful of joint-stock banks, each running a countrywide network of branches. This transformation began in earnest in the 1820s in Britain, and in the 1860s on the continent, reaching its present format by 1900.

The rise of deposit banking signaled a rise in financial agglomeration, for it triggered a trend toward branch banking. Joint-stock banks headquartered in the center threatened to turn local banks into local branches. Deposit banking also meant a concentration of the banking sector, though not necessarily the financial sector as a whole, as the consecutive expansion of the market spawned new functional specialization in the areas of issuing, brokerage, market making, and so forth. Branch banking did not develop uniformly across countries. Branch banking met the political opposition of local banks. In decentralized countries, local governments used their political power to stem branching, restrict deposit banking to the local economy, and stem the flight of local savings to the center.

Commercial banks varied between extreme specialization – as in Britain, where merchant banks, colonial banks, clearing banks, issuing banks, trustee banks, discounters, jobbers, stockbrokers, and savings banks each engaged in distinct financial activities – and universality – as in Germany, where all these activities were indifferently handled by large banks. More generally, a universal banking system is defined as a system in which banks engage in both lending and the underwriting of securities. In contrast, a specialized banking system keeps the two activities under separate roofs. Before 1913, universal banking is commonly held to have existed in Belgium, Germany, Austria, and Italy, whereas specialized banking was mostly encountered in France and the Anglo-Saxon countries.

Gerschenkron argued that universal banking reflected a median intensity in firms' capital needs – higher than in England, but lower than in Russia. From the perspective of the information asymmetry literature, universal banking is the most advanced form of delegated monitoring in conditions of acute information asymmetry.

The present account stresses the effect of capital spatial concentration.

The key variables of financial viability for any major project are costs (investment, financing, operations and maintenance) and revenues (mainly tolls in the case of transport projects). For each variable, forecast values may be different from actual values as documented above. Ar isk therefore exists that actual project viability may be substantially different from forecast viability.

The difference between forecast and actual viability may be so large that if the actual viability had been known for a given project, decision makers might have resolved:

1. (i) not to implement the project;

2. (ii) to implement the project in another form; or

3. (iii) implement another project.

In other words, non-viable projects, or projects that are less viable than forgone projects, may be implemented not because they are viable but because their viability was inaccurately predicted. The result would clearly be an inefficient use of resources.

Channel tunnel, Great Belt and Øresund

For the Channel tunnel, original estimates of viability have been rendered irrelevant by actual developments which have taken the project on a rollercoaster ride from expected high profitability to several near-bankruptcies. Most observers today consider the commercial viability of the Channel tunnel unproved and the prospect uncertain for original investors making a satisfactory profit.

After being issued at £3.50 per share on 9 December 1987, by mid-1989 initial optimism had spiralled Eurotunnel share prices to more than three times this value, above £11.00. Then delays and cost overruns hit the project, resulting in capital shortage and crisis.

With the necessary instruments for enforcing accountability elaborated in Chapter 10, we see two main alternatives for developing and reaching accountable decisions on whether or not to go ahead with a mega infrastructure project. The first alternative is based on the assumption that a project, if built, would be regulated by a concession, that is this alternative is based on the build-operate-transfer (BOT) approach, or a similar set-up. In the second alternative, a possible project, if built, would be constructed and operated by a state-owned enterprise (SOE). We assume in both alternatives that at least one third of the required capital will not be secured with government guarantees. We furthermore assume that the same taxation regime would apply to both alternatives.

The two main alternatives are set out in Tables XI.i and XI.ii, and should be compared to the conventional approach as summarised in Table VIII.i in Chapter 8. The first part of the process is similar for the two alternatives but differs towards the end, once all the public interest issues have been settled through the identification of performance criteria and regulatory provisions, and once it has been decided whether project implementation should be publicly or privately led. The steps involved in the development process would not necessarily follow each other in a simple consecutive order. Several of the steps would, most likely, be carried out simultaneously and with possible interactions and iterations.

This book is about moving money within and across countries. It raises the following question: are the few percentage points of my income that I save each month lent to a firm in my neighborhood or do they end up refinancing the short-term debt of the Republic of Mali instead? The answer to this question does not depend on technology, for, since the telegraph was invented, money has had the capacity to move to almost any urban area in the world at the speed of electromagnetic waves. Nor is the answer more likely to be found in economic reasoning. The local firm and the foreign government, holding risk constant, will pay the same interest on the sums they borrow. The answer, instead, is political. My savings are more likely to help fund production in my local industrial district if I live in Germany, Italy, Canada, or the United States, but to end up in Timbuktu if I live in Britain or France. Mobility of capital reflects the degree of centralization of the state. It is the structure of the state that determines the outreach of the “great go-between,” Bagehot's phrase for Britain's financial system, to which he ascribed the responsibility for moving money.

The second half of the nineteenth century saw the advent of modern finance. This is when the first joint-stock banks, financed by deposits rather than equity alone, opened their vast and sumptuous offices in London, Paris, New York, and other financial centers (chapter 3). This is when the first true international market for capital, backed by the most stable fixed-exchange-rate system ever to exist – the gold standard – emerged (chapter 4). This is when corporate stock markets opened their doors to the masses of individual investors (chapter 5), and when financial sectors across the industrializing world began to feel the need for product specialization (chapter 6). Except for automated tellers and internet connections, what we know of the financial sector today was already in place by the beginning of the twentieth century.

The previous chapter established that centralized countries registered relatively high levels of financial agglomeration. Without regulatory and legislative power, local governments in centralized countries could not – nor did they want to – stop center banks from opening a branch on every local “Main Street.” The reasoning, however, rested on an autarkic model – one without an international dimension. This chapter brings the international dimension into the picture, asking whether it ran parallel to or mitigated the process of agglomeration at work in the domestic economy. Theory and historical evidence both suggest that internationalization happened simultaneously with agglomeration, and more markedly in centralized than in decentralized countries.

Consider a country with two regions – the basic model of chapter 2. Graft onto it another country with a similar structure. Allow capital to flow freely between regions of a same country but not across countries. Burden the exchange of financial products with information asymmetry so that the losses incurred are lower between two financial cores than between a given core and a foreign periphery. The rationale for this differential in asymmetric information is that nineteenth-century foreign investors had an overwhelming preference for large, central, visible assets in foreign countries – mostly government bonds – over small, peripheral, and unfamiliar ones. It is reasonable to expect from such a model that the core–periphery pattern within a country and the degree of internationalization of the financial sector in that country would be mutually reinforcing.

In this and the following chapters we review the experience from a large number of megaprojects, including the Channel tunnel and the links across Great Belt and Øresund, all three multibillion-dollar projects. Although we subject these projects to critical scrutiny, our objective is not to criticise them, even where they have underperformed, but to learn constructively from experience by identifying lessons that may prove useful in improving future decisions regarding megaprojects. Given the large amounts of money spent on major transport infrastructure projects, it is remarkable how little data and research are available that would help answer the two basic questions: (i) whether such projects have the intended effects; and (ii) how the actual viability of such projects compares to projected viability. Therefore, in addition to transport projects, we have found it pertinent to review data and research from other types of infrastructure project and to compare experience from these projects with experience from the transport sector. In this manner, we will review data from several hundred large projects. In this chapter we focus on the costs of megaprojects. In Chapters 3 and 4 we consider the demand for, and viability of, such project.

The problem of cost overrun

Cost overruns in major transport infrastructure projects are widespread.

Finance is a rich field of study, pooling contributions from historians, political scientists, and economists. My goal is not to draw up an exhaustive inventory of the existing literature, but merely to situate the approach adopted in this book. I successively look at (1) the historical debate on capital scarcity, on which I offer a new perspective; (2) the use by political economists of the notion of capital mobility, which I try to clarify; and (3) the economic literature on information asymmetry, on which I build my argument.

Capital scarcity

In an article published in 1952, Gerschenkron provided the most ambitious explanation yet offered of why financial structures differ across nations. The more capital was needed in a short amount of time, he argued, the less equity markets could cope with the task of allocating long-term financial capital; instead, banks and state had to step in. Hence the “orderly system of graduated deviations from [the first] industrialization”: British industrialization was self- and market-financed, manufacturers ploughing back profits into their own factories; French industrialization (the 1850–70 spurt) was financed by investment bankers, who raised long-term capital and lent it to factories; German industrialization was financed by universal bankers, intermediating between depositors and factories; and Russian industrialization was financed by the state, raising capital from taxpayers and foreign lenders to distribute it to banks and factories. The need for banks or state intervention reflected economies of scale.