Economics in Times of Crisis

Cliometric Sessions at 1990 ASSA Meetings--December 28, 8:00 AM


Adam B. J. Klug
Princeton University

"We went too far in encouraging lending to Europe after the First World War. It turned out badly before we got through with it."

John H. Williams, testifying to Congress on the Marshall Plan in 1948. (Williams, 1952, 207)

"Deutschland kein Kolonialland ist."

Hjalmar Schacht (1927, 4)


In the later 1920s American investors became involved in a large and unprecendented amount of foreign lending. Between 1919 and 1929 U.S. long term investment abroad rose by $96 billion and accounted for about two-thirds of global new investment. This was quite different from the pre-war position in which Britain had been the major foreign investor and foreign creditor. By far the largest debtor to the United States was Germany, which furnished about a third of the new American assets. This situation was, however, not completely the outcome of the interplay of market forces, but was in reality the result of a policy, sometimes official, sometimes semi-official, which aimed at solving the pressing problems of war debts and reparations. As the two quotations above make clear, this policy was unpopular and controversial while it was actually being carried out. After Germany's default and the ruination of about 600,000 small American investors (Schuker, 1985, 357), it naturally incurred an odium which has endured to this day, and it is Keynes's sustained polemic of the 1920s which has remained lodged in the collective memory:

Practically the whole amount of the reparations so far have been provided by the foreign lender, and mainly by the United States. The United States lends money to Germany. Germany transfers its equivalent to the Allies, the Allies pay it back to the United States government. Nothing real passes - no one is a penny the worse off. The engraver's dies, the printer's frames are busier, but no one eats less, no one works more. (Keynes, 1978, 280-1)

Keynes, therefore, saw American lending as a device, and a futile one at that, for transferring reparation payments. Some recent views are even harsher in their evaluation of American lending. Thus, the distinguished historian Stephen Schuker writes that Germany not only did not pay but that

The governments of the Weimar Republic, during the successive stages of inflation, stabilization and deflation, sought to manipulate the international monetary system for Germany's benefit. In the short run they did so successfully. Not only did they avoid paying net reparations to the Allies, they actually extracted the equivalent of reparations from the former allied powers, and principally from the United States. (Schuker, 1985, 336)

Like Walter Benjamin's Angel, the historian's face is always turned to the past, and it is all too easy to reach a conclusion like Schuker's with the benefit of hindsight. John H. Williams, writing in 1952, presented a different view of events from the perspective of an influential economist who supported the policy at the time of its implementation:

The eventual losses have been an almost insurmountable barrier to further private investment. The conclusion, however, that our capital exports were mistaken is easier to reach now than it was at the time. The restoration of economic stability and of the gold standard and the large rebound in European production and trade that accompanied them in the last half of the 1920s, and in turn it was in that period that our capital exports were really large - were conditions calculated to invite investment which, in turn, stimulated production and trade. (Williams, 1952, 76)

Williams goes on to argue that the collapse of lending resulted from the Great Depression, whose severity was unprecedented and completely unforeseen. He stresses what a successful strategy German borrowing appeared to be at the time:

With these loans she was able to make her reparation payments under the Dawes Plan, to rationalise her industries and increase her capacity to pay. There was a body of respectable economic opinion which held that this was a logical way of solving the reparations problem so far as the German end of it was concerned, though it left unsettled the recovery of net remittances from Germany. (Williams, 1952, 77)

This attitude on the part of what Schumpeter terms "economic opinion" was summed up thus by two economists involved in the reparation debate:

The success with which the new loan policy was meeting led many to believe that the so-called transfer problem had been proved a myth. It was urged that inasmuch as the foreign loans were, in the main, extended for productive purposes, they would ultimately lead to an export surplus on the part of Germany and automatically provide the means for German payments on both reparation and private debt accounts. The fact that no difficulties had been found in procuring the foreign exchange with which to meet reparation instalments was cited in evidence, and it was argued that, thanks to the productive character of the loans, Germany was rapidly achieving a permanent export surplus, ample to meet both reparation and debt requirements. (Moulton and Pasvolsky, 1932, 381)

Williams, for example, expressed such sentiments when he stated that the "experience of the last five years" (Williams, 1930, 78) contradicted "those who magnify transfer difficulties" and that: "Never was theoretical expectation more completely and precisely confounded" (Williams, 1930, 73).

This paper re-examines this controversy from the 1920s. It should be pointed out that it does not deal with what is traditionally called the transfer problem. The issues raised in that context concern the so-called "secondary burden" of transfer payments which results from the terms of trade turning against the transferor country. Here, in contrast to the transfer problem literature, a one-good model with a single small country is used throughout. The issue addressed is that of the possibility of easing the "primary burden" of reparations by using borrowing to smooth consumption over time and finance investment. Thus this paper is an attempt to answer Charles Maier's (1979) request for clarification of the reparations problem from a dynamic perspective.


The convoluted proposals, counter-proposals and negotiations on the issues of Reparations have been exhaustively studied and reinterpreted by historians (for example, Schuker (1976), Trachtenberg (1979), Silverman (1982), Artaud (1979)). For our purposes the key point is that in May 1921 a debt in present value terms of 132 billion gold marks was imposed on Germany by the Allies' Reparations Commission. I calculate this to represent a debt of $500 per head of the German population at 1913 prices, allowing for approximately 40% inflation since 1914, and an exchange rate under the gold standard of 1 mark = $0.2382. This represents an enormous debt/GDP ratio of 3.1, where GDP is taken to be the average level of German output in the 1920s. For comparison, note that Mexico's debt/GDP ratio was 0.6 in 1983, while Brazil's was 0.4 (Solis and Zedillo, 1985), (Dornbusch, 1985). It should be pointed out, however, that the reparations bonds were divided into three categories, A, B and C. The A and B bonds had priority, and were less than half of the total. Even so, by any standards, the burden was considerable, and it is hardly surprising that the political and economic debate of the Ruhr occupation and German hyperinflation led to a major rethinking of the reparations question by all parties concerned. In November 1923 the Reparations Commission nominated two committees of non-political experts to examine various aspects of the problem and to propose solutions. The more important of the two, under the chairmanship of General Charles C. Dawes, a future Vice-President of the United States, was to consider means of stabilising German financial affairs and the possibilities for a new and more feasible schedule of reparations. This committee came up, in April 1924, with what became known as the Dawes Plan. For the purposes of this paper three aspects of the Dawes Plan should be noted. The first is that its ratification indirectly set off a spate of lending to Germany, a consequence which was envisaged by the Plan's framers. As recorded in Table 1 (available at the meeting), borrowing from the U.S. alone more than covered debt service on the reparations account.

A further intriguing provision of the plan was what was known as Transfer Protection. Under this device the German Government was to transfer marks to the Agent-General for Reparations who was responsible for converting them into foreign exchange for the Allies. He was not to do this, however, if the mark was consequently driven over the gold points, endangering the maintenance of the gold standard. This measure was regarded as sufficient both to deal with the transfer problem, and to buttress the monetary stability mentioned above. In practice, Germany's capital imports meant that there was never any lack of gold or foreign exchange and the clause was never invoked. The existence of this measure, however, justifies an analysis of German borrowing in 1924-9 predicated on the assumption that the transfer problem can be shunted on one side and the contemporary arguments of advocates of German borrowing, such as Williams, can be treated on their own merits.

The last aspect of the Dawes Plan which is important in what follows is the fact that the issue of the size of reparations debt was left open; the yearly payments alone were scaled down but 132 billion gold marks remained the nominal claim of the Allies, while a 'prosperity index', based on several indicators, such as the growth of exports, coal production and consumption of certain items was provided as a basis on which the size of the annual payments might be scaled up. Thus the German government could in effect influence the size of its debt by means of its economic decisions and had an incentive actually to restrict the growth of the economy. Furthermore, it is clear from the research of McNeil that neither the German, American nor British decision makers really believed that Germany would pay the full reparations debt (McNeil, 1986, 27).

Just how severe was the Great Depression in Germany is made completely clear by the second column of Table 1 (to be presented at the meeting) and against such a background it is hardly surprising that reparations payments were suspended.

The Dawes Plan did not end the history of reparations revision. In 1928 the American Agent-General for Reparations, Seymour Parker Gilbert, alarmed by the level of public expenditure in Germany and the process of "non-genuine transfers" of reparations facilitated by capital imports, succeeded, not without a little help from Schacht, in setting in train new attempts at reparations revision. This culminated in the Young Plan of 1929, which finally fixed the size of Germany's reparations debt, provided a new loan of about $300m and set up, for the first time, an international bank, the Bank for International Settlements, to service reparations payments. This plan never got off the ground and reparations were suspended in 1931. The final chapter in this tale was Schacht's suspension of much of the debt service in 1933, and his success in getting the European and British creditors to agree to a partial repudiation of the commercial debt. No agreement was ever reached with the United States about the private debts, and in the American case the German default was ultimately total.


In what follows I shall set out a simplified neo-classical model of optimal borrowing. The discussion procedes by means of stating and solving a variant of the model used by Cohen and Sachs (1986). Except for the assumption of a linear production function, this model is also very similar to the model used by Blanchard (1984) to study Brazil's debt problem, although it is less disaggregated.

3.1 Behaviour of firms The technology available to the country is described by a linear production function with one input, capital.

			Qt = aKt				(1)
where Q is GDP at time t, and K the capital stock. This capital depreciates at rate d so that the net increase in the capital stock is:

			\A\CO1(.,K, ) = It - dKt.
(2) In accordance with some remarks by Ohlin (1926), in his survey of the German economy in 1925, I assume that the installation of capital is costly. This also enables one to derive an expression for the investment function. These costs are related to total investment expenditure by the following formula

			J = (1 + \F(�,2) \F(It,Kt))It.				(3)
(where � is a parameter).
Firms try to maximise the discounted value of their profits, which simply equals (1) minus (3). It is shown in the appendix that their profits will be maximised by selecting a fixed rate of capital accumulation

			x = It/Kt.				(4)
The rate of growth of the economy simply equals this constant rate minus the rate of depreciation
		x - d = g = \A\CO1(.,K, )/K.	
It is also shown in the appendix that this rate of capital accumulation, x, depends on the interest rate and the technological parameters of the model.

3.2 Optimal paths By giving consumers the active role of maximising a utility function we can calculate the optimal path which gives them a preferred pattern of consumption over time. They maximise

		U = �\S( ,�, ,0, )e-dtU(ct)dt				(6)
subject to
 		�\S( ,�, ,0, )e-r(t)c(t)dt = � = -D(0) + �\S( ,�, ,0, ) q Ktc-rtdt
Here d is the subjective discount rate and � is total wealth. q is a-x(1+f/2). Wealth equals foreign asset holdings (-D(0), where D is debt) plus the present value of firms.

The standard solution to problem (6) is

		\F(\A\CO(.,c),ct) = b(r-d)
where b is the inverse of the elasticity of marginal utility.

If the utility function is isoelastic, and the world rate of interest is constant, consumption grows at a constant rate which is positive if the world interest rate exceeds the discount rate.

		ct = coeb(r-d)t					(8)

Substituting (8) in (7) yields.

		- \F(co,y-r) = -D(0) - \F(qKo,g-r)				(9)
Here y is the optimal rate of growth of consumption given by (7).

The appendix derives this result more fully, but it follows quite naturally from the separation of production and consumption decisions characteristic of neo-classical economics. In the appendix a result is derived for the value of the debt D(t), at later dates. The intuitive result holds that:

		D(t) = \F(ct,y-r)  -  \F(qKt,g-r)					(10)
where consumption and the capital stock equal coeyt and Koegt.

This is very similar to a standard debt accounting equation (e.g. Dornbusch (1988)) except that it is consistent with optimal behaviour over time and has a more complex paramaterisation.


We can now proceed to calculate optimal paths of debt accumulation and later a measure of solvency for the German economy starting in 1925, the year after the Dawes Plan was ratified. To some extent this exercise examines the validity of the policy embodied in the Plan, for whilst it did not provide directly for a stream of American capital exports to Germany, it was explicitly envisaged that the framework for stability, defined by the gold standard, Transfer Protection and the Agent-General's supervision would encourage them (Report of the Committee on Reparations, 1924, 20- 1). Of even more significance is the question as to whether the framers of the Young Plan, looking back on the experience of the previous five years, were correct in their assessments of the state of the German economy and as a counterpart to this, had American investors been rational in continuing to lend until the end of 1928? Explicit assessments will be given of the views of officials like Shepard Morgan, the director of the Economic Section of the office of the Agent-General, who claimed that:

"the foreign borrowing taken as a whole has more than paid its way ... at the close of the calculation the values existing in Germany, less the foreign debt incurred were larger than at the beginning. No observer of the economic progress Germany has made since 1924 can doubt that such in fact has been the case." (Shepard Morgan, 1930, 20)

4.1 Data

The rate of interest used is an average of Friedman and Schwarz's data on high grade U.S. corporate bonds in the relevant years. Depreciation is calculated from Balderston (1982). All other figures are taken from Hoffman (1965).

4.2 Closing the Model

The most useful characteristic of the model set out above is that it can be closed even with the limited data set of Table 8, included here. Adjustment costs can be calculated from the quadratic formula in the appendix, given that we know the rates of growth and depreciation. The identity between growth plus depreciation and the ratio of investment to the capital stock also allow us to sidestep the question of the reliability of Hoffman's investment data.

4.3 Tests for optimality and sensitivity analysis: Dawes Plan

This section calculates optimal paths of debt accumulation for the Weimar economy, given different assumptions about the size of the reparations debt.

The reader should remember that all debt projections are made under the counterfactual assumption of no Great Depression. They suggest what would have happened had "business as usual" continued after 1929. The initial level of consumption consistent with a particular debt burden emerges as a solution to the optimal programme. This can be compared with the initial consumption level chosen by economic actors in Germany and serves as a benchmark against which to judge the optimality of their behaviour. In addition, sensitivity tests are carried out to determine exactly what rates of growth of consumption and output would have been sufficient to meet the Allies' reparations demands without drastic reduction of German consumption.

Any level of debt between zero and $500 per head (RM132 billion) can be regarded as consistent with a possible optimal path. Zero is the upper bound, because, as Holtfrerich (1985) has shown, the hyperinflation liquidated all of Germany's foreign debt. $500 per head was the Allies' original reparations claim of 1922. Since there was no private debt in 1924 and historians believe that the Allies had concluded by then that $500 per head was their maximum claim (McNeil, 1986, 99), this figure is the upper bound.

The results of calculating initial values of optimal consumption, using equation (9), are contained in Table 3 (available at the meeting). The debt/GNP ratios in the table reflect reparations debt ranging from zero to the official figure of RM132 billion (i.e. a debt to GNP ratio of three) with an intermediate figure of RM60 billion.

The data, which are based largely on Hoffman's series, as explained above, suggest that German consumtion in 1925 was almost 2.4 times higher than that necessary to sustain a situation of zero foreign debt, let alone the payment of reparations with interest. Clearly Germany was a long way from any optimal consumption path.

I have carried out extensive sensitivity analyses on this model, some representative results of which are presented in Tables 4-6 (to be presented at the meeting). For example, Tables 5 and 6 show that had it been possible to return to the rate of growth of output of about 3% which prevailed in 1880-1913, this would have entailed reductions of about 35% in initial consumption. It should be pointed out that Table 5 shows that even maintaining an initial level of zero debt would have involved a sacrifice of consumption by Germany (about 6% of initial consumption), or alternatively a return to the rates of growth of output attained before the war. The tables also show that with no change in the growth rate, reparations payments would have implied a reduction of 10% in initial consumption per head and a fall in the optimal consumption growth from 1.5% per year to 1.0% per year.


The next set of calculation asks: what if the Great Depression and the Nazi seizure of power had not occurred? Would Germany have been able to afford the Young Plan obligations? These questions are of great historical interest, as it is generally believed that the Nationalist agitation against the exactions of the Young Plan was what provided Hitler with his springboard to power, and the model can shed some light on the potential effects of the Plan on German consumption (Craig, 1981, 526-8). Furthermore, Borchardt (1982) has recently created a stir by arguing that Germany's foreign obligations left her with so little room for manoeuvre that Bruning's fatal budgetary stringency in 1930-1932 was inevitable (Borchardt, 1982, 170-1). Again the model, by looking at the costs of the Young Plan to Germany, can shed light on this issue.

The Young Plan itself involved the fixing of a final figure for the Reparations debt, the abandonment of transfer protection and the dismantlement of the Agent-General's office. From the economic point of view what matters is that the German economy was left with a debt of $140 per head in real terms according to my calculations, instead of an unspecified debt with an upper limit of $500, as was the case under the Dawes Plan. $140 at 1913 prices per head represents the Young obligations of 37 billion Reichsmarks plus the commercial debt Germany had built up in the interim. I now examine the implications of the acceptance of the Young Plan by the Reichstag and the Referendum of 1930.

The actual growth rates for 1925-9 generate immediate surpluses of about 12 billion R.M., in real terms, six times the forecast for surpluses made by the Germans to the Young Committee (James. 1985, 75). However, this implies extremely low levels of initial consumption of about $70 per head, increasing output or reducing the rate of growth of consumption yield more realistic paths for the current account which are almost exactly in line with what was projected at the Paris Conference. With a growth rate of 3.0% per annum, Germany ceases to be a net debtor after 18 years while, if consumption grows at 2.4% and the growth rate is 2.3% this process takes 23 years. All this can only be achieved at a very high cost, unrealistic falls of about 40% in the initial level of consumption. Those scenarios which maintain initial consumption at roughly the 1929 level of $183 per head result only in steeply rising current account deficits. The scenario of the second column on the left whereby consumption falls by about 15% is more realistic, but this is still greater than the 11% fall wich actually took place in 1929-1932. These results point to the conclusion that the Young Plan, if it was to lead to a German balance-of- payments surplus, necessarily involved a drastic reduction of living standards. A government such as Bruning's, concerned for its foreign credit position and determined to honour its reparation obligations would have had, according to the model, little choice but to cut public consumption and raise taxes. The hypothesis put forward by Fleisig (1974), that Bruning's motivation for meeting the Young schedule was to show concretely that Germany's task was impossible politically, is also supported by the draconian results with regard to the necessary reduction in consumption.

An acute business economist and observer of the international economic scene, with a confessed prediliction for equilibrium-type theories, Benjamin Anderson, was therefore correct when he stated that the Young Plan depended on Germany "reversing radically the whole course of economic life" and in particular reducing consumption (Anderson, 1945, 207).


Up until this point my methodology has been based on the calculation of optimal paths under various assumptions, comparing them with historical experience, It now seems natural to start assessing the validity of the views of contemporaries like James W. Angell that "the prospects for German default or repudiation" were "negligibly small" (1930, 88).

There exist two possibilities for applying the existing theoretical literature to this above issue. One is to use Cohen's (1986) solvency index, and therefore to make the same calculations for Germany in the 1920s as he has made for developing countries in the 1980s. The other possibility is to exploit the fact that one can calculate at least the official costs of default in the German case, as these were set out in the Dawes Plan. It is, therefore, possible to asses whether these sanctions were sufficient to deter default. If they were indeed sufficient, we can conclude that Angell was correct to dismiss this possibility.

6.1 Were the Default Penalties Sufficient?

I now present a model of international borrowing which is even more simplified than that of section 5, since it assumes there is no investment and output simply grows at a constant exogenously given rate.

In this case the country's rulers maximize:

		�\S( ,�, ,0, )e-dt(1n ct)dt
		s.t.   \A\CO1(.,D, ) = rDt  +  Q0egt				(11)
		Dt  given, \A\CO1( ,lim,t��) e-rtDt = 0.

where d is the rate of time preference.
Within this framework, based on the work of Eaton and Gersovitz (1981), it is assumed that a country may repudiate its debt if the cost of repaying it becomes "excessive". If a country does so its creditors immediately drive it into financial autarky and some actual financial penalties are paid by the defaulter. The country, therefore, has to chose between the benefits of remaining within the international loan market and the net gains from default, given the autarkic utility level which can be achieved with the domestic capital to which the country has access after it has defaulted on its debt. This autarkic level is defined as follows. Take a country which defaults at time t*. Assume that its resources are reduced by a factor ls-t at time s, and that it is forced into financial autarky after time t*. In this case a country defaulting at t* may receive the endowments (1 - ls-t*)Qs at time s, where Qt is output and equals aKt. The default penalty ls-t* need not be a constant and may become heavier or higher over time. At each point of time t therefore, the country has the ability to get an autarkic level of utility after default defined by:

		Ua(Qt)  =  �\S( ,�, ,t, )e-btln(cs)ds				
			s.t.  cs  =  (1 - ls-t)Qs.			(12)
Call U(Qt, Dt) the utility the country would obtain by not defaulting at time t. This level of utility is derived from the solution of (7.1). The problem facing lenders is to set the default penalty lt at a level which ensures that we always have Ua(Qt) � Ut(Dt,Qt) for all t. In an appendix it is proved that the solution to this problem implies that

		d  �\S( ,�, ,t, )  e -d(s-t) 1n(1-ls-t)ds  =  1n (\F(ct,Qt)).	(13)
and that this in turn implies that the debt ceiling Dt is determined by

		Dt  =  - \F(lQt,(g-r))					(14)
Since the growth rate is constant in this model, this formula is close to one advocated by Kemmerer, the chief economic advisor to the Dawes Committee, by which annual reparations payments would be a constant proportion of total output (see Dawes, 1939), as opposed to the use of the Prosperity Index.

The default penalty is now found by examining the provisions against default made under the Dawes Plan. Their legal framework is explained in Moulton's book The Reparations Plan, (Moulton 1924, 151), while their economic costs can be gleaned from Angell's account, (Angell 1929, 67) and from Auld's (Auld 1927, 201-9). They placed half the capital of the German railways in the form of bonds whose yearly interest would be paid directly to the Allies as part (about 40%) of the Reparations payments. The railways themselves were placed under the control of a French Railway Commissioner, who could take over the railways completely in the event of default. A further five billion marks worth of commercial and industrial debentures were also placed at the disposal of the Agent-General. In the event of a German default the Allies would, therefore, be able to enjoy half the capital income of the German railways, and, on the basis of Hoffman's data on the value of German companies, about 31% of the capital income of German business and industry. In 1925 this comes to only 1.9% of GDP, although in 1929 the returns to capital of the railways were one third higher than in 1925 and the returns in industry were more than twice as high. Taking an average of the returns in the corporate sector and in the railways over 1925-29, the potential default penalties are still only 1.9% of GDP.

The results of using these figures to calculate the debt ceiling using equation (14) are recorded in Table 10 (to be presented). Recall that the debt ceiling is that level of debt which maintains the ratio of debt to the present value of output equal to the default penalty, and this is that level which keeps the utility of continuing repayment equal to the utility achievable under autarky. At any higher level of debt, the default penalties will be insufficient to deter the country from defaulting and choosing autarky. Making this calculation yields a debt ceiling of $162 per head or a debt output ratio of 0.95. This is the maximum level of reparations debt which would have made it worthwhile not to repudiate, given the default penalties. Note, however, that this is higher than the figure of $140 per head calculated in the previous section which represents Germany's total debt under the Young Plan plus her commercial debt in 1929. Unfortunately, the Young Plan abolished the default provisions.

These results show that the default penalties were sufficient to force Germany to pay a moderate level of reparations. Clearly they would also have prevented default on the smaller commercial loans. There is, however, a great deal of uncertainty as to whether these penalties covered the commercial loans or not. One contemporary observer and adviser of American bankers, Benjamin Anderson, apparently did think that the default penalties would be used for this purpose, although the Dawes Plan stated no such thing explicitly (Anderson 1946, 120). The initial Dawes Loan, however, was covered by these penalties and there is evidence that lenders believed this measure would be repeated in other cases (Huertas and Cleveland 1985, 154-60). McNeil has shown in addition that American lenders did try to have the default penalties extended to their loans (McNeill, 1987, 73-4, 40-2).

6.2 The Cohen Solvency Index

Daniel Cohen (1985), uses the above model differently to derive an index of solvency. Taking (14) as his starting point he calculates the fixed fraction of resources which should be devoted to servicing the debt. His framework is slightly different because he assumes discrete time and that lending begins one period ofter the initial one, but essentially there is no difference. Under these assumptions debt service will equal:

		b  =  \F(r-g,1+g) \F(Dt,Qt).					(15)

where b is the fraction of output devoted to debt service, and b � l, where 
l is the default penalty.
The important concession to realism in Cohen's approach lies in the fact that he does not assume that the economy is initially in a steady state, so that (15) only holds in year T, several years after the initial year of the calculation. Between years t and T the growth and interest rates can fluctuate from year to year. The analogue to equation (14) in this case is therefore:

		D0 = b \I\SU(,,)\S(T,t =1) \F(Qt,P\S(t,1)(1+ri)) + \I\SU(,,)\S(T,t 
=1) \F(DT,P\S(T,i =1)(1+ri))).	(16)

	From (15) we find that:

		DT = (\F(r-g,1+g)) \F(1,b) QT.				(17)

	Substituting this expression in (5.8) and using the fact that 
QT = \I\SU(,,)\S(T,1)Q0(1+yt), yields the following expression for the 
resources devoted to debt service:
		b =\B\BC\[(\A\CO1( , , ,\I\SU(,,)\A\CO1(T,1) \F(P\S(T,1) 
(1+gi),P\S(t,1) (1+ri))   +)  \F(P\S(T,1) (1+gi),\F(P\S(T,1) (1+ri),(\F(r-
g,1+g))))) \A\CO1( , ,-1, , , , , , , , ,). \F(D0,Q0) .	(18)
In order to use (18) in a practical calculation one must decide on a value of T. In Cohen's calculations this was taken to be ten years after the initial year, following which the economy settles down to a steady state with constant interest and growth rates. In this case I take the initial year as 1924, and the final year as 1929. Continuing the counterfactual methodology of my previous discussion, after 1929 the economy is assumed to follow the average growth and interest rates which prevailed in 1925-29. In other words this is a solvency index with the Great Depression left out. Again it is a measure of the realism of expectations in the 1920s, given that contemporaries involved in negotiating the Young Plan like Thomas Lamont and analysing the Weimar economy like Angell stress that a Depression of such magnitude was completely unexpected, and that once underway, an upturn of the business cycle was regarded as just around the corner (Angell, 1932, 372), (Lamont, 1930, 94).

The following table shows values of the solvency index which is the proportion of GNP or exports needed to produce a trade surplus which would enable the debt to be serviced, given that the country must remain solvent. Cohen also uses exports as an appropriate measure of resources.


Total Reparations		6.5%		37%
A and B bonds only		2.5%		14%

The index of solvency is used by Cohen in the following way: he compares the solvency index - in effect an indication of what should be repaid, to the amount the country actually did pay in the form of a trade surplus. The data in Tables 1 and 2 in section 2 make it clear enough that Germany was very much in deficit in 1925-9. Unsurprisingly, Germany did not make the required adjustment to debt repayment. This is true even if one considers the commercial debt. This was reduced to only $3 million by the hyperinflation, but Germany was not a net creditor even when one assumes the elimination of reparations. Therefore she should have had at least a balanced trade account on average during the 1920s, even when reparations are not taken into account. In one year, 1926, there was a small surplus, but this was only 2.6% of exports and only 1.2% of GNP. On the other hand, under the Bruning government, Germany did attempt an adjustment, and in 1931 the surplus was 7.9% of exports and only 1.5% of GNP. This shows, interestingly, that the adjustment achieved by Bruning was sufficient to render Germany solvent under the terms of the Young Plan.

How much of an adjustment would have been required from Germany in contemporary terms? In other words, how large a trade surplus did she have to create? Cohen calculated a value of 13% of exports for his solvency index for Latin America in 1984. This figure can be compared with the results given here in Table 7. Clearly the index suggests a very great adjustment, even by contemporary standards, if the whole reparations debt of 1922 was demanded. The A and B bonds alone or the slightly lower demands of the Young Plan debt would have involved a degree of adjustment comparable to that required in Latin America in the 1980s.


The analysis in the previous sections raises serious questions about the wisdom of the actions of the negotiators of the Dawes and Young Plans and the wisdom of American investors in lending to Germany. So we face the crux of the problem - why did American experts support the reparations demands of the Dawes and Young Plans and why did American loans flow to an insolvent Germany? My answer to this question is a historian's answer rather than an economist's. That is to say, my explanation is based on an assessment of the situation and knowledge facing policymakers in the 1920s, not on economic criteria. In the contingency of that particular time and place, appropriate concepts were lacking in order to analyse the problem. The point is that economic knowledge grows and transmutes itself over time and that one's interpretation of economic decisions must take the state of knowledge and competence at that time into account.

7.1 The Productive Credit View

The illusion that Germany could pay was sustained by an influential theory held by some economists and well received by economic opinion. This theory, emphasised the consumption augmenting effects of productive investment, financed by borrowing in the U.S. and it was believed that these would enable both increased consumption and the payment of reparations (Angell, 1929, 335), (Brown, 1930, 90), (Williams, 1930, 78). These writers criticised both the budgetary view (e.g. Graham, 1925) that a "steady volume of payments abroad more or less automatically, given monetary and budgetary precautions, gives rise to conditions which will make the transfer a permanent thing" (Angell, 1926) and the competing transfer view to the effect that German export surplus was "rigid" (Moulton and McGuire, 1923, Chapter 3; Keynes, 1922, Chapter 4). This criticism was of the view's "dateless quality" (Williams, 1930, 70), while the new school's presumptive theoretical case takes account of borrowings as a constructive part of the reparations process" (Williams, 1930, 78), which "would rob the debate between the other two points of view of its significance" (Brown, 1930, 90).

I have termed this theory, following the terminology of the time, "the productive credit view" (Von Sering, 1928, 201). The model of equations (1-9) can serve as a rational reconstruction and encapsulation of this school of thought. Surprisingly one can find discussions of dynamic planning by firms (Angell, 1929, 213), a linear production function (Moulton, 1925, 208), the implications of an infinite horizon, (Hawtrey, 1932, 110) and intertemporal solvency (Auld, 1928, 184). Even intertemporal optimisation in the Fisherian model was treated by F. A. Fetter in the context of borrowing to repair war devastation (Fetter, 1926, 103-4).

The productive credit view began as a justification for the Dawes Plan (Young, 1924, Auld, 1927) and appeared with greater vigour as the intellectual basis for the Young Plan (Williams, 1930, 77), (Angell, 1929, 330-5). This view even had footholds in Britain, (Guilleabaud, 1924) and even in Germany, (Weber, 1928), despite the objections of the Reichsbank's advisor (Bonn, 1928, 154) and Schacht himself (1926, 4-5, 14-15). The impact of Angell's Recovery of Germany (1929) can be traced in Stimpson's notebooks (Link, 1970, 400-1), of that of Auld in the notes of the German Foreign Ministry's reparations expert Karl Ritter (Akten zur deutschen ausw�rtigen Politik, series B, Band XVIII, No. 74) and of William's writings on the German Economics Ministry, (document reproduced in Maurer and Wengst 1980, 606-10).

Yet this approach had a crucial weakness: a lack of what an economic assistant to the Dawes Committee later termed a lack of "economic intelligence", i.e. "accurate information penetratingly interpreted" (Davis, 1975, 304). To highlight this point, I have constructed a set of "pseudo- data" presented here in Table 8, reflecting the actual beliefs about the German economy held by members of the productive credit school. These data can be compared and contrasted with those currently accepted by economic historians by means of the table shown here. The main, but not exclusive source of these incorrect estimates is Angell's Recovery of Germany (1929) supposedly "a sound piece of workmanship" (Economic Journal Review, March 1930), and an "exceptionally well-balanced piece of work" (Sir A. McFadyean, Economic Journal, March 1931), demonstrating "a thorough exploitation of statistical sources" (AER review by M. Hartshough, May 1930). At least in part the appearance of "the most remarkable economic recovery in the history of the World" (Angell, 1930, 81), derives from the special circumstances of the recovery from hyperinflation in 1924.

By feeding these data into the models set out above, the consistency of the views of the productive credit view can be judged. The results are summarised here in Table 9. Whatever initial value of debt/GNP is chosen between 0 and 3.5, all produce initial values of consumption within the range stated or implied by this contemporary literature. Neglecting repudiation risk explicitly (Angell, 1932, 177), was not a major failing, since Germany remains solvent with these "notional" data. Thus the Young Plan in particular appears to have been an honest mistake since the "notional" data and Table 8 suggest that Germany could have met her solvency criteria under the Young Plan, and in the first-best world of optimal borrowing, would have had her welfare substantially increased.


.c.;			Notional Data		Actual Data
			(1924-1928)		(1925-1929)
a. annual rate
    growth of NNP 		7.1%						2.4%
b. annual rate of
     growth of capital 
     stock		6.2%						2.0%
c. annual rate of
    growth of
    consumtion		1.9% (low estimate)	
		3.6% (high estimate)				4.1%

d. GDP deflator		-						1.4%

e. wholesale price
     index		-2%						-

f. rate of interest		8.5%						4.4%

g. depreciation		7.0%						4.0%

h. capital stock
    per head, 1924		$612						$877 

i. aggregate consumption
    per head, 1924		$106 (low estimate)
		$139 (high estimate)				$163 (1925)

Sources: 	Notional data
		a,b,e,f. Angell (1929)
		c,i.     Moulton and McGuire (1924), Angell (1929)
		g.     Thomas (1934)
		h.     Moulton and McGuire (1924), Stamp (1930)

		Actual data
		Hoffman (1965) except g. calculated from Balderston (1982)
All growth rates are geometric.


Test				Actual Data		Notional Data
Optimal path adhered to 
with perfect capital markets	not optimal		optimal

Default penalty deters default	Default on A and B	No default on A 
and B
(i.e. size of the default penalty	bonds is best strategy	bonds, or on 70% 
ensures a time-consistent					the C bonds
Cohen solvency index compared	twice as insolvent as	As solvent as 
 to its value for debtors 		Latin America in 1984	America in 
of the 1980s			(With debt consisting of	(With debt 
				A and B bonds only, 	consisting of A and
				German solvency in 	B bonds only
				1924=Latin America 	German is close to
				1984)			complete solvency).


What implications do these findings have for some of the judgements historians have made about the issues of reparations and German commercial borrowing in the 1920s? On the whole they point to a middle way between those like Costigliola (1984), who laud the constructive nature of American policy and those like Schuker (1985) and Marks (1978) who regard Germany both as being able to pay reparations and as consistently planning to default on her commercial obligations. The first view appears to be incorrect because American policy was bedevilled by a severe lack of "economic intelligence". On the other hand the commercial loans and the A and B bonds would have been protected by the default penalties, had the political will existed to impose them. Even so it has not been proven that Germany could not have easily created current account surpluses. Even under the Young Plan schedule this would have involved reductions of 30-40% in the initial level of consumption. The picture of overall weakness in the low-growth Weimar economy, painted by Borchardt (1982) and James (1985) certainly translates into a formal model, since the growth rates of output and consumption actually prevailing in 1925-29 clearly result either in behaviour well off the optimal growth path in a model with perfect capital markets or in insolvency in the case of a model with repudiation risk.