Confidence and alternative Keynesian methods of asset choice

Abstract

This article elaborates the analysis of asset choice proposed by Keynes and later adopted by Post Keynesians such as Paul Davidson and Hyman Minsky. The article incorporates the essential aspects of the theory of confidence presented in Dequech (1999), first into an investigation of the relation between confidence and the liquidity premium and then into the broader theory of asset choice. Keynes considered two methods of determining planned investment expenditures: one method is based on the comparison between the marginal efficiency of capital and the interest rate; the other is based on the comparison between the demand price and the supply price of a particular capital good. Both methods can be generalized for asset choice, with investment as a particular case. The article refines and develops these two methods so as to specify more precisely the influence of confidence and speculation on the determination of liquidity premia and hence on the several assets' profitability.

Acknowledgments

I would like to thank two anonymous referees for extremely valuable comments and suggestions. Thanks are also in order to André Orléan and Olivier Favereau, who welcomed me as a visiting researcher at Cepremap-École Normale Supérieure and Forum-Université de Paris 10, where this article was completed. Financial support from the National Research Council (CNPq) and the CAPES Foundation of the Brazilian Ministry of Education is also gratefully acknowledged.

Notes

¹See Dequech (2003) for a more detailed discussion of Keynes's ideas in formal terms, including several points of notational clarification and an analysis of the conditions for expected profit maximization and for equivalence between the two methods.

²This conception of confidence especially applies to situations of fundamental uncertainty, in which the list of all possible relevant events is not predetermined. In contrast, the notion of ambiguity, which can also be used in combination with the notion of confidence, goes beyond the mainstream conception of uncertainty (or risk), but still falls short of fundamental uncertainty. Under ambiguity, people are uncertain about probabilities but the list of all possible relevant events is predetermined and usually known (Dequech, 2000a). On the factors affecting the degree of uncertainty, see Dequech (2004).

³Keynes (1936, pp. 240n, 148n) also related the liquidity premium to the notion of weight he had presented in A Treatise on Probability, which reinforces the connection established in Dequech (1999) between confidence and weight. The link between liquidity premium and weight returns in a 1938 letter to Townshend (Keynes, 1979, p. 293); see also O'Donnell (1989) and Runde (1994).

⁴The term ‘speculator’ is being used here in a strict sense, to denominate someone who believes himself to know ‘better than the market what the future will bring forth’ (Keynes, 1936, p. 170) and, more specifically, to be able to ‘outwit the crowd’ (p. 155), ‘to guess better than the crowd how the crowd will behave’ (p. 157). Keynes also referred to speculation in a broader sense, as ‘the activity of forecasting the psychology of the market’ (p. 158). It is suggested here that one can perform this activity by conventionally following the current average opinion or by speculating against the latter, i.e., speculating in the strict sense.

⁵When allowing innovations to happen, we must take an investment decision to mean not only a quantitative decision to increase capacity (as in Keynes, 1936, with the hypothesis of a given technique), but also a qualitative decision as to which type of capital good must be bought—or produced by the investing firm itself. Implementing an innovation may require, or consist of, introducing a new type of capital good, made according to the specifications of the innovator.

⁶Katzner (1998) illustrates the difficulties involved here by discussing a function in which a cardinally measured variable such as production depends on ordinally gauged effort. The case of equation 4 is more complex in that there are four arguments instead of one, but it is simpler in that none of the arguments is a choice variable and hence no maximization problem is involved. Relatedly, the signs of the second-order derivatives are not very relevant.

⁷Recent developments in investment theory have also incorporated the desire for flexibility (as the desire to avoid irreversibility) in an intertemporal setting with weak uncertainty; see, for example, Dixit & Pindyck (1994).

⁸In this sense, an expected negative A _j is different from an expected decrease in Q _j  − C _j . This is why Q _j and C _j are not arguments in the L _i function.

⁹The less liquid the asset i is, the less its possessor can take advantage of the expected negative appreciation of other assets; money, for example, can be used for this speculative purpose because, when the time comes, it can be easily liquidated for the asset j whose price is expected to fall. The less liquid the asset j is, the less the expectation of its negative appreciation will attract speculators; if the asset j is a capital good, for example, people will tend not to speculate with it, because, even if its price is expected to go up again after going down, it cannot be easily sold. In other words, as Davidson (1978, pp. 198–199) notes, the question of whether one is a bear or a bull depends on the transaction costs of buying and selling an asset.

¹⁰Waiting, in this case, is different from waiting in the argument that associates liquidity preference with the desire for flexibility due the possibility of learning and revising one's choice. In the former case, people wait for something that they are confident will happen; in the latter, they wait because they expect to learn more about what will or may happen or about that which may increase their confidence.

¹¹The assumption of a representative capital good avoids the difficulties involved in the possibility of using a different capitalization rate for different capital goods and, more generally, for different assets. This question is discussed in a separate paper. It is assumed here that the cash flows Q, C, L and A are net of any discount specific to a particular asset, over and above the general discounting through the liquidity premium of money.

¹²Although the appreciation that contingencies exist could be likened to a perception of uncertainty, it is not obvious that this appreciation should vary in degree, let alone be measured cardinally. Moreover, even if Minsky wanted this appreciation to express something like a degree of uncertainty, it is not clear why his ‘state of uncertainty’ includes ‘an evaluation of the alternative outcomes’, which could mean in a different context exactly that which one is uncertain about, rather than uncertainty itself.

¹³For Keynes (1936, p. 144), in contrast, what he somewhat confusingly calls the borrower's risk is, ‘if a man is venturing his own money, … the only risk which is relevant’. Keynes's alternative expression for it, ‘the entrepreneur's risk’, is more generally appropriate.

¹⁴Although confidence considerations could be said to lead to diversification in the specific sense that a portion of the portfolio would tend to be held in the form of at least one highly liquid asset, the diversification issue that Minsky discussed under the label of borrower's risk could be solved by diversifying only the types of illiquid assets held, and thus is not strictly related to liquidity.

¹⁵The present value of an illiquid asset has to be properly reduced to compensate for the loss of flexibility. This argument is incorporated through the influence of uncertainty perception on the rate of discount and is the other side of the argument that liquidity may be preferred because it allows flexibility, just as capital goods are, in relation to money, on the other side of the liquidity spectrum. A similar idea, with a weaker notion of uncertainty, has been introduced to the neoclassical literature on investment and irreversibility (e.g., Dixit & Pindyck, 1994).