Wiley 3fe The Case for Financial Laissez-Faire Author(s): Kevin Dowd Source: The Economic Journal, Vol. 106, No. 436 (May, 1996), pp. 679-687 Published by: Wiley on behalf of the Royal Economic Society Stable URL: http://www.jstor.org/stable/2235576 Accessed: 15/09/2013 08:04 Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at http://www.jstor.org/page/info/about/policies/terms.jsp JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org. 9 STOR Wiley and Royal Economic Society are collaborating with JSTOR to digitize, preserve and extend access to The Economic Journal. http://www.jstor.org This content downloaded from 147.251.194.197 on Sun, 15 Sep 2013 08:04:37 AM All use subject to JSTOR Terms and Conditions The Economic Journal, 106 [May), 679—687. © Royal Economic Society 1996. Published by Blackwell Publishers, 108 Cowley Road, Oxford OX4 iJF, UK and 238 Main Street, Cambridge, MA 02142, USA. THE CASE FOR FINANCIAL LAISSEZ-FAIRE Kevin Dowd The argument for financial laissez-faire (or free banking) is essentially very simple: if free trade is generally desirable, then what is wrong with free trade in the financial services sector? If nothing is wrong with it, the whole panoply of government intervention into the financial sector — the central bank, government-sponsored deposit insurance and government regulation of the financial system - should presumably all be abolished. If there is something wrong with laissez-faire, on the other hand, then what exactly is the problem with it? Why does this problem justify intervention? And why does it justify the particular interventions we have, such as a central bank? Most economists take a patently untenable position on these issues. For the most part, they accept the general principle of free trade, but they deny that it applies to financial services. Yet relatively few could give a coherent defence of this position or have even thought that much about it. They oppose free banking more or less instinctively, as if its failings are obvious. The response, of course, is that what is obvious is not necessarily true - the history of science is full of cases where the 'obvious' turned out to be wrong. It therefore behoves us, as academic economists, to explore these issues more carefully and beware of assuming we already know the answers before we start. Before getting into detail, I would like to make three general points. First, if free trade is good, as most of us agree, there must be at least a prima facie case in favour of free banking. If the principle of free trade applies in general, we must presume it to apply to any specific individual case, unless we have clear reason to believe the contrary. The onus of proof is on those who oppose free banking to demonstrate its undesirability. Most professional economists have the wrong priors on this issue. Secondly, while I accept that free banking seems strange at first sight, I believe this reaction mainly reflects the way we have been conditioned to think. Our education leads us to take certain things for granted, and the need for central banking is one of them. After all, why else do we initially react so strongly to what is no more than the application of the generally accepted doctrine of free trade to financial services? Finally, there is a great deal of empirical evidence on the free banking issue, and this evidence is supportive of the predictions of free banking theory, and, in particular, of its claim that unregulated banking is stable. Economists cannot therefore maintain that free banking has never been tried, nor that it has been tried and 'failed'. The evidence also supports the predictions made by free banking theory that intervention generally weakens the financial system and causes the very problems it is ostensibly meant to cure. By contrast, the evidence is also inconsistent with opposing theories that have been suggested as providing justifications for central banking. [ 679 ] This content downloaded from 147.251.194.197 on Sun, 15 Sep 2013 08:04:37 AM All use subject to JSTOR Terms and Conditions 68o the economic journal [may I. THE FREE BANKING POSITION A Laissez-Faire Financial System So what would a free banking system look like, and how would it operate? Imagine a hypothetical laissez-faire economy with an underlying 'imperfect' economic environment — information is scarce and asymmetric, there are non-trivial agency and coordination problems, and so on. These problems give rise to a financial system characterised by the presence of intermediaries that enable agents to achieve superior outcomes to those they could otherwise achieve (e.g. by cutting down on transactions and monitoring costs). Perhaps the most important intermediaries are banks. These invest funds on behalf of client investors, some of whom hold the bank's debt liabilities and others its equity. Most bank debts are deposits of one form or another, and most of these can be redeemed on demand. Many deposits are also used to make payments by cheque. The equityholders are residual claimants, and their capital provides a buffer that enables a bank to absorb losses and still be able to pay its debtholders in full. The banking industry exhibits extensive economies of scale, but not natural monopoly,1 and there is typically a small number of nationwide branch banks, with a larger number of specialist banks that cater to niche markets. The industry is competitive and efficient by any reasonable standard. We can think of this banking system as operating on a convertible, commodity-based monetary standard (e.g. a gold standard).2 Bank liabilities are denominated in terms of the economy's unit of account (e.g. the pound), and underlying the system is some rule that ties the unit of account to a unit of the 'anchor' commodity on which the monetary standard is based.3 The price level in this system is then determined by conditions in the market for the 'anchor' commodity.4 Bank currency is convertible - the banks must redeem their currency when required to — and so the amount of currency in circulation 1 There is much evidence of economies of scale in banking, but no evidence that these economies of scale are so large that the industry is a natural monopoly. It follows that one cannot defend the central bank's monopoly privileges over the currency supply on the grounds that free banking would lead to a currency monopoly anyway. Nor should natural monopoly be confused with the use of a single economy-wide unit of account. There will typically be one generally used unit of account (e.g., the pound), but the use of a single unit of account reflects economics of standardisation (or economies in use) and not natural monopoly, which necessarily involves economies of production. See Dowd (1993), ch. 5. 2 The alternative is to assume that we have a free banking system based on convertibility into a frozen monetary base (as discussed, e.g., by Selgin, 1994). It seems to me that a commodity-based system is more natural: all historical free banking systems were of this type, and the assumption of a frozen monetary base presupposes some earlier government intervention - an assumption that can be awkward if one is trying to assess the validity of intervention in the first place. 3 For example, in a gold standard, the rule would require that banks of issue peg the exchange rate between their currency and gold (or, if one prefers, peg the nominal price of gold). We can then think of the equilibrium nominal prices of individual commodities as determined by the combination of relative demand and supply factors and the fixed nominal price of gold. Gold can thus be regarded as the ' anchor' commodity that ties down nominal prices throughout the economy. 4 Suppose there is a gold discovery under a gold standard. At initial prices, there is now an excess supply of gold and excess demands for other commodities. The relative price of gold against other commodities must therefore fall, but the nominal price of gold cannot adjust because the rules of the gold standard hold it fixed. Hence, the relative price of gold falls by means of a rise in the nominal prices of other goods. We can therefore think of the price level as determined by the factors that determine the relative price of gold, and hence by demand and supply in the gold market. © Royal Economic Society 1996 This content downloaded from 147.251.194.197 on Sun, 15 Sep 2013 08:04:37 AM All use subject to JSTOR Terms and Conditions 1996] case for financial laissez-faire 68l is determined by the demand to hold it. If banks issue too much currency, the public simply return it to the banks for redemption, and the excess currency is automatically retired. The Stability of Financial Laissez-Faire But how stable is the system? With no lender of last resort or state-run deposit insurance system, depositors would be acutely aware that they stood to lose their deposits if their bank failed. They would therefore want reassurance that their funds were safe and would soon close their accounts if they felt there was any significant danger of their bank failing. Bank managers would understand that their long-term survival depended on their ability to retain their depositors' confidence. They would therefore pursue conservative lending policies, submit themselves to outside scrutiny, and publish audited accounts. They would also provide reassurance by maintaining adequate capital. The greater a bank's capitalisation, the more losses a bank can withstand and still be able to pay off depositors in full. If the bank's capital is large enough — if the bank is adequately capitalised — the bank can absorb any plausible losses and still repay depositors, and depositors can be confident their funds are safe. The precise amount of capital is determined by market forces. The better capitalised a bank is, other things being equal, the safer and more attractive it is to depositors, but capital is also costly, and depositors need to pay shareholders to provide it (e.g. by accepting lower interest on deposits). Competition between banks should then ensure that banks converge on whatever degrees of capitalisation their customers demand (and, by implication, are willing to pay for): banks will be exactly as safe as their customers demand. If bank customers want safe banks, market forces will ensure they get them (see, e.g. Dowd J994*)- The conclusion that banks under laissez-faire would maintain high levels of capital is also consistent with the empirical evidence (see also Benston and Kaufman (BK), 1995, pp. 3-4)- For example, US banks in the antebellum period were subject to virtually no federal regulations and yet had capital ratios in most years of over 40% (Kaufman 1992, p. 386). US banks were subject to more regulation at the turn of the century, but even then their capital ratios were close to 20%, and capital ratios were still around 15% when federal deposit insurance was established in 1933-4 (Kaufman, 1992). The evidence is also consistent with the associated prediction that laissez-faire banks face low probabilities of failure. Thus US banks appear to have been fairly safe in the period before the Civil War (Dowd, 1992, ch. 11), and for the period afterwards, Benston et al. (1986), pp. 53-9 report that bank failure rates were lower than the failure rates for non-financial firms. Losses to depositors were correspondingly low (Kaufman, 1988). Failure rates and losses were also low for other relatively unregulated systems such as those in Canada, Scotland, Switzerland and various others (see, e.g., the case studies in Dowd (1992)). Nor is there any reason to expect banking instability arising from the ways in which banks relate to each other, either because of competitive pressures, or because of' contagion' from weak banks to strong ones. It is frequently argued © Royal Economic Society 1996 This content downloaded from 147.251.194.197 on Sun, 15 Sep 2013 08:04:37 AM All use subject to JSTOR Terms and Conditions 682 the economic journal [may that competitive pressures produce instability by forcing 'good' banks to go along with the policies of 'bad' ones (e.g. Goodhart 1988, pp. 47—9). The underlying argument seems to be that if the bad banks expand rapidly, they can make easy short-term profits which pressure the managers of good banks to expand rapidly as well, with the result that the banking system as a whole cycles excessively from boom to bust and back again. However, a major problem with this argument is that it is not in a bank's interest to engage in aggressive expansion of the sort this argument envisages. A bank can usually expand rapidly only by allowing the average quality of its loans to deteriorate, and a major deterioration in its loan quality will undermine its long-run financial health and its ability to maintain customer confidence. A profit-maximising bank will not choose to undermine itself this way, even if other banks appear to be doing so. Indeed, if a bank believes that its competitors are taking excessive risks, the most rational course of action is for it to distance itself from them - perhaps to build up its financial strength further — in anticipation of the time when they start to suffer losses and lose confidence. The bank is then strongly placed to win over their customers and increase its market share at their expense, and perhaps even drive them out of business. The bank would have to forgo short-term profits, but it would win out in the long run. There is no reason, then, to suppose that competitive pressures as such would force free banks into excessive cycling.5 Then there is the contagion argument that the difficulties of one bank might induce the public to withdraw funds from other banks and threaten the stability of the financial system. The conclusion normally drawn from this argument is that we need a central bank to prevent 'contagion' by providing lender of last resort support to a bank in difficulties. However, this argument ignores the earlier point that good banks have a strong incentive to distance themselves from bad ones. If the good banks felt there was any danger of contagion, they would take appropriate action — they would strengthen themselves and curtail credit to weak banks — to help ensure that contagion did not in fact occur. Indeed, as discussed already, they would position themselves to offer the customers of weaker banks a safe haven when their own banks get into difficulties. A serious danger of contagion is therefore inconsistent with equilibrium. Instead, we would expect the difficulties of a weak bank to trigger a 'flight to quality' in which customers transfer their accounts to stronger banks, and this expectations is borne out by the evidence which tells us that runs occur in response to news that a particular bank or group of banks has sustained major losses that call into question its ability to repay its debts (see, e.g., Kaufman, 1988). When runs occur, the typical scenario is a flight to quality, with substantial inflows of funds to the stronger banks, and there is little evidence that runs are contagious (see, e.g., Benston etal. 1986, pp. 53—60; 5 Proponents of the excessive cycling theory sometimes look to examples such as the excessive bank lending to LDCs in the late 70s and early 80s (e.g. Goodhart (1988, pp. 48-g), Dow (1996, pp. 700, 702)). However, episodes like these are not examples of free banking and can hardly be held up as examples of what would happen under it. Many national authorities were actively encouraging their banks to make loans to LDCs, and banks could reasonably expect some form of bailout if things went bad. In the circumstances, it was therefore hardly surprising that they over-reached themselves. See Dowd (1994a) p. 306. © Royal Economic Society 1996 This content downloaded from 147.251.194.197 on Sun, 15 Sep 2013 08:04:37 AM All use subject to JSTOR Terms and Conditions 1996] case for financial laissez-faire 683 Benston and Kaufman, 1994; Dowd, 1994«, p. 297). The contagion hypothesis is thus implausible and empirically rejected. The Impact of State Intervention There is also the issue of what happens to this system if the government intervenes in it. There is no space to consider here all the ways in which governments intervene, but two specific interventions are particularly important — the establishment of a central bank to provide lender of last resort (LLR) support to the financial system, and the establishment of a state-sponsored system of deposit insurance - both intended, according to their proponents, to stabilise the banking system. The establishment of an LLR is meant to provide liquidity to banks that cannot otherwise obtain it. Since good banks can always obtain loans to maintain their liquidity, an LLR therefore protects bad banks from the consequences of their own actions. It therefore directly encourages the very behaviour — greater risk-taking and the maintenance of weaker capital positions — that a sound banking regime should avoid. It also undermines the discipline of the market in another, less obvious way. Since an LLR in effect tries to keep weaker banks open, its very existence reduces the incentives for good banks to adopt the virtuous strategy of building themselves up in anticipation of winning weaker banks' market share. That strategy depends on the weaker banks facing ruin and cannot promise much payoff if the LLR is going to bail them out. In these circumstances, even the (otherwise) good banks may decide to take greater risks and let themselves go. Ironically, the LLR can then produce the very instability that proponents of central banking often claim would arise under free banking. In fact, someone who observed this instability might easily attribute it to the market itself, and falsely believe that the banking system actually needs the LLR support that is, in reality, undermining it. A major cause of banking instability — the LLR — could easily be mistaken for its cure - and, unfortunately, often is. A system of deposit insurance has comparable effects. Once we introduce deposit insurance, depositors no longer have any incentives to monitor bank management and managers no longer need to worry about maintaining confidence. A bank's rational response is to reduce its capital, since the main point of maintaining capital strength — to maintain depositor confidence - no longer applies (see, e.g., Dowd 1994^). Even if an individual bank wished to maintain its capital strength, it would be out-competed by competitors who cut their capital ratios to reduce their costs and passed some of the benefits to depositors by offering them higher interest rates. The fight for market share would then force the good banks to imitate the bad. Deposit insurance consequently transforms a strong capital position into a competitive liability, reduces institutions' financial health and makes them more likely to fail. It also encourages more bank risk-taking at the margin: if a bank takes more risks and the risks pay off, then it keeps the additional profits; but if the risks do not pay off, part of the cost is passed on to the deposit insurer. The bank therefore takes more risks and becomes even weaker than its capital ratio alone would suggest. These claims are borne out by the evidence: the claim that banks reduce © Royal Economic Society 1996 This content downloaded from 147.251.194.197 on Sun, 15 Sep 2013 08:04:37 AM All use subject to JSTOR Terms and Conditions 684 the economic journal [may their capital ratios is confirmed by the observation that US bank capital ratios more than halved in the 10 years after the establishment of federal deposit insurance (e.g. Kaufman, 1991), and there is abundant evidence that US deposit insurance has increased failure rates and associated losses. II. BENSTON AND KAUFMAN Benston and Kaufman (BK) agree with much of this analysis, but disagree on the central bank LLR function and on government deposit insurance. On the former issue, they argue that a central bank LLR is necessary to provide liquidity assistance and avoid banks with liquidity needs from having to sell assets at 'firesale losses' (e.g. Kaufman, 1991, p. 106). However, I would object that this position depends on the unjustified assumption that free banking would not provide adequate liquidity. If a bank does not have good collateral, we agree that it should not get assistance; but if it does have good collateral, it presumably can get loans or liquidity provided other banks are free to provide them, and other banks would grant those loans if the terms were right; hence, there should be no need under free banking to establish a central bank to provide lender of last resort assistance. If banks cannot provide such assistance, it must presumably be because of legal restrictions, in which case the first-best solution is not to establish a lender of last resort, but to abolish the legal restrictions. Nor is a central bank-LLR as a perfect substitute for the automatic support mechanisms of the free market. The LLR creates serious moral hazard problems, as well as a need for system management and 'policy' decisions (and all their attendant problems) that does not arise under free banking. The record of historical central bank LLR operations does not inspire confidence either, as when much of the US banking system collapsed in the early 1930s because of the Fed's refusal to grant the very support it was established to provide. BK have no convincing answer to these problems, and I fail to understand why they insist on trying to patch up a patently unsatisfactory system when a much better one is available. BK support government deposit insurance for a number of reasons. In large part, they support it because they believe that the alternative of private deposit insurance is seriously inadequate (e.g. because of the limited capacity or credibility of private insurers; cf. Benston et al. 1986, p. 83; BK, 1996, p. 693), but they also support it because they believe there would be serious external effects in its absence, such as the danger of a run to currency (e.g. BK, 1994, pp. 4—5). However, I believe these arguments miss the point. The relevant choice is not that between government and private deposit insurance, but that between government deposit insurance and the free-banking scenario outlined earlier, in which banks provide implicit insurance through strong capitalisation. If bank customers demand safe banks — as we all agree they do - the free market would provide safe banks via appropriate bank capitalisation, not through private deposit insurance. The reason is straightforward: providing depositor security through deposit insurance creates conflicts of interest between bank shareholders and the insurance agency, and handling these conflicts is costly; © Royal Economic Society 1996 This content downloaded from 147.251.194.197 on Sun, 15 Sep 2013 08:04:37 AM All use subject to JSTOR Terms and Conditions 1996] case for financial laissez-faire 685 however, providing security through bank capitalisation does not create such conflicts, and is therefore a more efficient means of ensuring depositor safety. Private deposit insurance would of course be allowed under free banking, but we would not expect to see it, and this expectation is confirmed by the evidence. I readily accept their point that bank failures involve third-party effects, but would argue these effects merely reinforce my position. The third-party-effects argument is essentially an argument for the safer system, and we have already seen that the safer system is free banking. It is therefore an argument in favour of free banking and against deposit insurance. In any case, the danger of third-party effects is not unique to bank failures, and applies to the failures of most other institutions as well. So BK must argue either that bank failures are 'special' in ways that other failures are not, or they must be consistent and advocate a policy of preventing all institutional failures, and not just those of banks. Yet they are unwilling to do either. III. SHEILA DOW Sheila Dow advances two principal arguments against free banking. The first is a version of the earlier argument that free banking produces excessive cycling (e.g. Dow, 1993, p. 199; 1996, p. 702), from which she concludes that there is a need for the state or a central bank to 'stand above' the market process and take action to correct the excessive cycling to which laissez-faire is prone. Her argument is distinctive in that it rests on an underlying post-Keynesian view of the uncertainty attached to valuing bank assets and not just on a herd theory of bankers following each other over the cliff. However, despite its distinctiveness, it is still open to the objections already made against the excessive cycling theory: it ignores the point that individual free banks do have incentives to go against the market trend, and thereby counteract it, because doing so enables them to increase their long-run market shares; and it is empirically falsified, because there is little evidence of such cycles under historical free banking systems. In any case, even if free banking does produce some cycling, it still does not follow that it produces excessive cycling to which some form of intervention is an appropriate solution. Intervention has costs of its own - it uses up resources, the parties involved have their own interests, and so on; it cannot therefore be treated as a costless process that is guaranteed to produce the result its advocates hope for. One must also keep in mind that the historical record suggests that real-world intervention has destabilised the financial system rather than stabilised it (e.g. Dowd (ed.), 1992). Dow's post-Keynesian version of the cycling argument also has its own distinctive problems. If assets are so difficult to value, then how can we expect central bankers to know where the private bankers are going off the rails? If they don't know, how can we expect them to ameliorate cycles and not inadvertently add to them? And even if they do know, what can they do about it if they are as powerless over the cost and volume of credit as Dow seems to suggest they are?6 6 i also reject her view that the argument for free banking hinges on the ignoring of uncertainty. Uncertainty is intrinsic to economic life, and banks have to live with it as much as anyone else. The critical issue is not whether uncertainty exists, but whether uncertainty makes valuation as difficult as she suggests © Royal Economic Society 1996 This content downloaded from 147.251.194.197 on Sun, 15 Sep 2013 08:04:37 AM All use subject to JSTOR Terms and Conditions 686 the economic journal [may Dow also opposes free banking for another reason. She effectively argues that free banking is pointless because central banking would emerge spontaneously from it (e.g. Dow and Smithin, 1994, p. 21), or at least would do so in cases where the free banking was 'successful' and did not lead to some form of chaos (e.g. Dow, 1996, p. 704). The Dow-Smithin (DS) argument is that the requirement of a credit-based economy produce a degree of centralisation of power that is effectively the same as central banking. When a crisis occurs under free banking, there is a scramble for safe assets, a scramble that focuses on the banking system's ultimate settlement asset. The issuer of this asset then has great power over the system and becomes de facto a central bank (see, e.g. Dow, 1996, pp. 704; DS, 1994, pp. 14—5). To give this argument empirical support, DS point to the role of the three big banks in making emergency loans, disciplining minor banks, and the like, in early nineteenth century Scottish free banking (e.g. DS, 1994, p. 25). There are a number of problems with this argument. (1) While it is true that crises do produce a scramble for safe assets, we would expect this scramble to take the form of a flight to quality, and the evidence confirms that it generally does; it does not usually take the form of a scramble for some ultimate settlement asset, whatever that might be. There tends to be recycling within the banking system, rather than a flight from it. (2) Even if there was a scramble for ' ultimate' settlement assets, it still does not follow that such assets are issued by one institution only. To assume that there is only one key institution consequently begs the point at issue. One must also keep in mind that the experience of historical free banking suggests that there is always more than one 'big' bank; no bank ever spontaneously established its supremacy over all the others. To assume that there is only one key institution in the free banking system therefore also flies in the face of the evidence. (3) Even if one accepts their logic, DS establish the existence of the wrong animal. At most they establish the possibility of a private bankers' 'club', a club that might impose 'rules' on and sometimes assist its members, but establishing the existence of full-blown central banking is an altogether different matter. The powers of a private bankers' club would be (and historically were) extremely limited: membership would be voluntary in a way that 'official' regulations are not; club powers would be determined by the banks themselves, not by some outside agency, and as a corollary, club officials would be accountable to the banks rather than to outsiders; and, perhaps most important, the scope of club rules and the demands they would make on members would be far less than those implied by modern central banking (see, e.g., Dowd, 1994«). (4) And, finally, even if one accepts their whole argument, they still have no substantial objection to free banking. If they are right, they merely establish that free banking would be pointless, but not harmful; but if I am right, then free banking is best and central banking is harmful. But whoever is right, we would it does. In my view, she exaggerates these difficulties, and I can only agree with BK (1996, pp. 691) when they say that banks and the public do in fact find ways of valuing bank assets and liabilities that work reasonably well in practice. © Royal Economic Society 1996 This content downloaded from 147.251.194.197 on Sun, 15 Sep 2013 08:04:37 AM All use subject to JSTOR Terms and Conditions 1996] case for financial laissez-faire 687 never choose to suppress free banking in favour of central banking — free banking wins by Pascal's wager. IV. CONCLUSIONS None of the critics of free banking has presented a convincing case that free banking is an exception to the general rule in favour of free trade. The prima facie case in favour of free banking therefore still stands. However, one should also assess the competing theories against the evidence, and the evidence clearly indicates that historical free banking was in fact stable, just as free-banking theory predicts. The conventional view that free banking could not be stable must therefore be rejected - and with it the conventional wisdom that tries to justify central banking and state intervention in the financial system. Sheffield Hallam University References Benston, George,J., Eisenbeis, R. A., Horwitz, P. M., Kane, E.J. and Kaufman, G. G. (1986). Perspectives on Safe and Sound Banking: Past, Present, and Future. Cambridge, MA: MIT Press. Benston, G.J. and Kaufman, G. G. (1988). Risk and Solvency Regulation of Depository Institutions: Past Policies and Current Options. New York: Salomon Brothers Center, Graduate School of Business, New York University, Monograph 1988—1. Benston, G. J. and Kaufman, G. G. (1994).' Is the banking and payments system fragile?' Journal of Financial Services Research (forthcoming). Benston, G.J. and Kaufman, G. G. (1996). 'The appropriate role of banking regulation'. Economic Journal, this issue, pp. 688-697. Dow, S. C. (1993). Money and the Economic Process. Aldershot: Edward Elgar. Dow, S. C. (1996). 'Why the banking system should be regulated.' Economic Journal, this issue, pp. 698-707. Dow, S. C. and Smithin, J. (1993). 'Change in financial markets and the "first principles" of monetary economics.' Mimeo, University of Stirling, and York University, Ontario. Dowd, K. (1992). (ed.) The Experience of Free Banking. London: Routledge. Dowd, K. (1993). Laissez-Faire Banking. London: Routledge. Dowd, K. (1994a). 'Competitive banking, bankers' clubs, and bank regulation.' Journal of Money, Credit, and Banking, vol. 26, no 2 (May), pp. 289-308. Dowd, K. (19944). 'Banking, costly liquidation, and deposit insurance.' Mimeo, School of Financial Studies and Law, Sheffield Hallam University. Goodhart, C. (1988). The Evolution of Central Banks. Cambridge, MA: MIT Press. Kaufman, G. G. (1988). 'Bank runs: causes, benefits, and costs.' Cato Journal, vol. 7, no. 3 (Winter), pp. 559-87- Kaufman, G. G. (1991). 'Lender of last resort in contemporary perspective.' Journal of Financial Services Research, vol. 5, no. 2 (October), pp. 95-110. Kaufman, G. G. (1992). ' Capital in banking: past, present, and future.' Journal of Financial Services Research, vol. 5, no. 4 (April), pp. 385-402. Selgin, G. A. (1994). 'Free banking and monetary control', Economic Journal, vol. 104, no. 427 (November), pp. 1449-59. © Royal Economic Society 1996 This content downloaded from 147.251.194.197 on Sun, 15 Sep 2013 08:04:37 AM All use subject to JSTOR Terms and Conditions