Abstract
We study relationships between different aspects of risk preferences. We show that, under the assumptions of non-satiation and bounded marginal utility, some additional conditions on the asymptotic behaviour of the indices of relative prudence and relative temperance ensure that risk aversion, prudence and temperance are equivalent. Similar conclusions are derived for higher-degree risk aversion. Moreover, some links between indices of relative risk aversion of different degrees are derived. The implications of these results for several economic problems which involve risk changes are discussed.
Similar content being viewed by others
Notes
De Donno et al. (2019) derived, in the context of a portfolio problem, some restrictive conditions which ensure that the threshold condition on the index of partial relative risk aversion of some order implies the threshold conditions on the same indices of lower orders. We also recall that Maggi et al. (2006) studied the linkages between the monotonicity of the index of relative risk aversion and the index of relative prudence.
These are just some of the possible applications of our results, which can generally be used in all problems where changes in risks are relevant, such as production localization problems (as in Eeckhoudt et al. 2009) labour supply choice (as in Chiu and Eeckhoudt 2010), optimal self protection (as in Crainich et al. 2016).
Throughout the paper, we say that the limit of a function f as x approaches infinity exists when \(\lim _{x\rightarrow +\infty } f(x) = l \in {\mathbb {R}}\) or \(\lim _{x\rightarrow +\infty } f(x) = \pm \infty\)
Consider as an example for the risk averse case the classical CRRA utility and for the risk lover case the utility function \(U(x)=x\;\arctan (x)-\log (1+x^2)\) on the domain \([0, +\infty )\).
Many economic models suggest that this additional utility should decrease when wealth increases. This is what is called the “Law of diminishing marginal utility” (e.g. Marshall 1920, p.79). This idea, however, implies risk aversion, which is clearly an assumption we do not introduce in our paper since risk aversion is one of the aspects of preferences involved in our results.
All rational functions composed of exponentials, logarithmic, polynomial functions or the inverse tangent have a limit at \(+\infty\) (which can be infinite).
References
Barro, R., & Sala-i-Martin, X. (2004). Economic Growth (2nd ed.). New York: McGraw-Hill.
Brockett, P. L., & Golden, L. L. (1987). A class of utility functions containing all the common utility functions. Management Science, 33, 955–964.
Caballé, J., & Pomanski, A. (1996). Mixed risk aversion. Journal of Economic Theory, 71, 485–513.
Chiu, W. H. (2005). Skewness preference, risk aversion, and the precedence relations on stochastic changes. Management Science, 51, 1816–1828.
Chiu, W. H., & Eeckhoudt, L. (2010). The effects of stochastic wages and non-labor income on labor supply: update and extensions. Journal of Economics, 100, 69–83.
Chiu, W. H., Eeckhoudt, L., & Rey, B. (2012). On relative and partial risk attitudes: theory and implications. Economic Theory, 50, 151–167.
Crainich, D., Eeckhoudt, L., & Hammitt, J. K. (2015). The value of risk reduction: new tools for an old problem. Theory and Decision, 79, 403–413.
Crainich, D., Eeckhoudt, L., & Menegatti, M. (2016). Changing risks and optimal effort. Journal of Economic Behavior & Organization, 125, 97–106.
De Donno, M., Magnani, M., & Menegatti, M. (2019). Changes in multiplicative risks and optimal portfolio choice: new interpretations and results. Decisions in Economics and Finance. https://doi.org/10.1007/s10203-019-00250-1.
Denuit, M. M., & Eeckhoudt, L. (2016). Risk aversion, prudence, and asset allocation: a review and some new developments. Theory and Decision, 80, 227–243.
Eeckhoudt, L., & Gollier, C. (2005). The impact of prudence on optimal prevention. Economic Theory, 26, 989–994.
Eeckhoudt, L., Gollier, C., & Schlesinger, H. (1996). Changes in background risk and risk-taking behavior. Econometrica, 64, 683–90.
Eeckhoudt, L., & Schlesinger, H. (2008). Changes in risk and the demand for saving. Journal of Monetary Economics, 55, 1329–1336.
Eeckhoudt, L., Schlesinger, H., & Tsetlin, I. (2009). Apportioning of risks via stochastic dominance. Journal of Economic Theory, 144, 994–1003.
Ekern, S. (1980). Increasing \(N\)th degree risk. Economic Letters, 6, 329–333.
Gollier, C., & Pratt, J. W. (1996). Risk Vulnerability and the Tempering Effect of Background Risk. Econometrica, 64, 1109–1123.
Kimball, M. S. (1990). Precautionary saving in the small and in the large. Econometrica, 58, 53–73.
Kramkov, D., & Schachermayer, W. (1999). The asymptotic elasticity of utility functions and optimal investment in incomplete markets. Annals of Applied Probability, 9, 904–950.
Leland, H. E. (1968). Saving and uncertainty: the precautionary demand for saving. Quarterly Journal of Economics, 82, 465–473.
Levy, H. (2006). Stochastic Dominance: Investment Decision Making Under Uncertainty (2nd ed.). New York: Springer.
Maggi, M., Magnani, U., & Menegatti, M. (2006). On the relationship between absolute prudence and absolute risk aversion. Decisions in Economics and Finance, 29(2), 155–160.
Marshall, A. (1920). Principles of economics (Revised ed.). London: Macmillan and Co.
Mas-Colell, A., Winston, M. D., & Green, J. (1995). Microeconomic Theory. Oxford: Oxford University Press.
Menegatti, M. (2001). On the conditions for precautionary saving. Journal of Economic Theory, 98, 189–193.
Menegatti, M. (2009). Optimal prevention and prudence in a two-period model. Mathematical Social Sciences, 58, 393–397.
Menegatti, M. (2014). New results on the relationship among risk aversion, prudence and temperance. European Journal of Operational Research, 232, 613–617.
Menegatti, M. (2015). New results on high-order risk changes. European Journal of Operational Research, 243, 678–681.
Menezes, C. F., Geiss, C., & Tressler, J. (1980). Increasing downside risk. American Economic Review, 70, 921–932.
Menezes, C. F., & Wang, X. H. (2005). Increasing outer risk. Journal of Mathematical Economics, 41, 875–886.
Romer, D. (2012). Advanced Macroeconomics (4th ed.). New York: McGraw-Hill.
Rothschild, M., & Stiglitz, J. E. (1970). Increasing risk: I. A definition. Journal of Economic Theory, 2, 225–243.
Rothschild, M., & Stiglitz, J. E. (1971). Increasing risk II: Its economic consequences. Journal of Economic Theory, 3, 66–84.
Sandmo, A. (1971). On the theory of the competitive firm under price uncertainty. The American Economic Review, 61, 65–73.
Schachermayer, W. (2002). Optimal investment in incomplete financial markets. In H. Geman, D. Madan, S.R. Pliska, T. Vorst (Eds.) Mathematical Finance—Bachelier Congress 2000 (pp. 427–462). Berlin, Heidelberg: Springer Finance, Springer.
Scott, R. C., & Horvath, P. A. (1980). On the direction of preferences for moments of higher order than the variance. The Journal of Finance, 35, 915–919.
Treich, N. (2010). Risk-aversion and prudence in rent-seeking games. Public Choice, 145, 339–349.
Acknowledgements
We would like to thank the Associate Editor and two anonymous referees, whose comments and suggestions helped to improve the presentation of our results.
Author information
Authors and Affiliations
Corresponding author
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Appendix: Lemmas and proofs
Appendix: Lemmas and proofs
Lemma 1
Under non-satiation, if there exists\(x_0 \ge a\)such that\(U^{\prime \prime }(x) < 0\)for all\(x\in [x_0, +\infty )\), then
- (i)
\(\lim _{x \rightarrow +\infty } U^{\prime }(x) = l\ge 0\)
- (ii)
\(\lim _{x \rightarrow +\infty } U^{\prime \prime }(x)=0.\)
- (iii)
If \(\lim _{x \rightarrow +\infty } xU^{\prime \prime }(x)\)exists, then it must be 0.
Proof
To prove (i), we observe that since \(U^{\prime \prime }(x) <0\), the marginal utility \(U^{\prime }(x)\) is strictly decreasing, it has a limit as \(x \rightarrow +\infty\). Being \(U^{\prime }(x) > 0\), the limit must be greater than or equal to 0.
Since \(U^{\prime }(x) = U^{\prime }(x_0) + \int _{x_0}^x U^{\prime \prime }(t)\mathrm{d}t,\) and \(\lim _{x \rightarrow +\infty } U^{\prime }(x) = l < + \infty\), the integral \(\int _{x_0}^{+\infty } U^{\prime \prime }(t)\mathrm{d}t\) converges. Since \(U^{\prime \prime }(x)\) is constant in sign, we have necessarily \(\displaystyle \lim _{x \rightarrow +\infty } U^{\prime \prime }(x)=0,\) so (ii) is proved.
To show (iii), note that the function \(xU^{\prime \prime }(x)\) is strictly negative, so the limit, if it exists, is some \(k \le 0\) (k can possibly take the value \(-\infty\)). Assume by contradiction that \(k<0\). Thus, there exist some \(M < 0\) and \(z \ge \max (x_1, 0)\) such that \(xU^{\prime \prime }(x) \le M\) on \([z, +\infty )\), or equivalently \(U^{\prime \prime }(x) \le \frac{M}{x}\). Integrating between z and x for \(x > z\), we obtain the following inequalities:
where the first inequality follows from the fact that \(U^{\prime }(x)\) is decreasing and converges to l when \(x \rightarrow + \infty\). We thus obtain a contradiction. \(\square\)
Lemma 2
Under non-satiation and marginal utility bounded at\(+\infty\), if there exists\(x_0 \ge a\)such that\(U^{\prime \prime \prime }(x) > 0\, \forall x\in [x_0, +\infty )\), then, in addition to\(U^{\prime \prime }(x) < 0\)and conditions (i) and(ii) of Lemma 1we have that
- (iii)
\(\displaystyle \lim _{x \rightarrow +\infty } xU^{\prime \prime }(x)=0.\)
- (iv)
\(\displaystyle \lim _{x \rightarrow +\infty } U^{\prime \prime \prime }(x)=0.\)
- (v)
\(\displaystyle \lim _{x \rightarrow +\infty } xU^{\prime \prime \prime }(x)=0.\)
Proof
We have that for \(x> y > 0\),
Exploiting the mean value theorem and the fact that \(U^{\prime \prime }\) is increasing, we obtain the following relations, with \(\xi \in (y,x)\):
All the quantities above are negative, because \(U^{\prime \prime }\) is negative (and \(U^{\prime }\) decreasing). Taking the absolute values, the inequality becomes \(|U^{\prime \prime }(x)(x-y)| \le | U^{\prime }(x) - U^{\prime }(y) |\); hence we obtain
Claim (iii) thus follows from (i) and (ii) in Lemma 1.
The convergence of \(U^{\prime \prime }(x) = U^{\prime \prime }(x_0) + \int _{x_0}^{x} U^{\prime \prime \prime }(t) \mathrm{d}t\) as \(x \rightarrow + \infty\) implies that the integral \(\int _{x_0}^{+\infty } U^{\prime \prime \prime }(t) \mathrm{d}t\) is finite; hence \(\lim _{x \rightarrow +\infty } U^{\prime \prime \prime }(x) = 0\) (since \(U^{\prime \prime \prime }(x) > 0\)) and (iv) is proved.
To prove claim (v), we first observe that by the integration by parts formula we have
On the other hand \(\int _{x_0}^x U^{\prime \prime }(t) \mathrm{d}t = U^{\prime }(x) - U^{\prime }(x_0)\). Combining these two equalities, we obtain
We then apply the same argument as in the proof of claim (iv). \(\square\)
Lemma 3
Under non-satiation and marginal utility bounded at\(+\infty\), if there exists\(x_0 \ge a\)such that\((-1)^{n-1} U^{(n)} (x) > 0\, \forall x\in [x_0, +\infty )\), then we have that
- (vi)
\(\displaystyle \lim _{x \rightarrow +\infty } U^{(j)}(x) =0\) for \(j=2, \ldots , n\)
- (vii)
\(\displaystyle \lim _{x \rightarrow +\infty } xU^{(j)}(x)= 0\) for \(j=2, \ldots , n\).
Proof
Proposition 3 in Menegatti (2015) implies that \((-1)^{j-1} U^{(j)} (x) > 0\) for \(j=1, \ldots , n\) for all \(x\in [x_0, +\infty )\). Then (vi) and (vii) can be proved by induction, exploiting the same reasoning as in the Proof of (iv) and (v) in Lemma 2. \(\square\)
Proof of Theorem 1
-
(a)
The implication \(\mathrm{(ii)}\, \Rightarrow \, \mathrm{(i)}\) follows from Menegatti (2014), Proposition 1(b), with \(x_1=x_2\).
To prove the opposite implication, we first show that \(\mathrm{ARP}(U) \ge 0\). Assume by contradiction that \(\mathrm{ARP}(U)< 0\). Then there should exist a value \({\bar{x}} > 0\) such that \(-\frac{x U^{\prime \prime \prime }(x)}{U^{\prime \prime }(x)} < 0\) for all \(x\ge {\bar{x}}\). It follows that \(U^{\prime \prime \prime }(x) < 0\) for all \(x\ge \max ({\bar{x}}, x_1)\). But, in this case, Proposition 1(a) in Menegatti (2014) implies \(U^{\prime \prime }(x) > 0\) for all \(x\ge \max ({\bar{x}}, x_1)\); hence we obtain a contradiction.
Assume now that \(\mathrm{ARP}(U)= k,\) with \(0\le k < + \infty\). Then \(\left| \frac{x U^{\prime \prime \prime }(x)}{U^{\prime \prime }(x)} \right|\) is definitely bounded by some constant \(C>0\). In addition, we know that \(U^{\prime }(x)\) tends to a finite limit l (Lemma 1(i)). If we fix \(\varepsilon > 0\), the two previous conditions allow us to find \(z \ge x_1\) such that \(\left| \frac{x U^{\prime \prime \prime }(x)}{U^{\prime \prime }(x)} \right| \le C\) and \(|U^{\prime }(x)-U^{\prime }(y)| \le \varepsilon\) for all \(x,y \in [z, + \infty )\). Moreover, as in the proof of Lemma 2, we can write \(\int _x^y U^{\prime \prime }(t) \mathrm{d}t = xU^{\prime \prime }(x)-yU^{\prime \prime }(y) - \int _y^x t U^{\prime \prime \prime }(t) \mathrm{d}t\). Then, for \(x\ge y \ge z\)
$$\begin{aligned} |xU^{\prime \prime }(x)-yU^{\prime \prime }(y)|= & {} \left| \int _y^x \left( U^{\prime \prime }(t) + t U^{\prime \prime \prime }(t) \right) \mathrm{d}t \right| \le \int _y^x\left| U^{\prime \prime }(t) \left( 1+\frac{tU^{\prime \prime \prime }(t)}{U^{\prime \prime }(t)}\right) \right| \mathrm{d}t \\\le & {} (1 + C) \int _y^x\left| U^{\prime \prime }(t)\right| \mathrm{d}t = (1+C) |U^{\prime }(x)-U^{\prime }(y)| \le (1+C)\varepsilon . \end{aligned}$$This implies that \(\lim _{x \rightarrow +\infty } xU^{\prime \prime }(x)\) exists, and by Lemma 1 (iii), it is equal to 0. We can thus apply L’Hôpital’s rule to \(-\frac{xU^{\prime \prime }(x)}{U^{\prime }(x)-l}\) and obtain
$$\begin{aligned} 0 \le \lim _{x\rightarrow +\infty } -\frac{xU^{\prime \prime }(x)}{U^{\prime }(x)-l} = \lim _{x\rightarrow +\infty } \left( -1 -\frac{x U^{\prime \prime \prime }(x)}{U^{\prime \prime }(x)} \right) = -1+k. \end{aligned}$$Hence, either \(1 \le \mathrm{ARP}(U) < + \infty\) or \(\mathrm{ARP}(U) = +\infty .\) In both cases, for a fixed positive constant c, one can find some \(x_2 \ge x_1\) such that \(\displaystyle -\frac{x U^{\prime \prime \prime }(x)}{U^{\prime \prime }(x)} \ge c\), that is
$$\begin{aligned} U^{\prime \prime \prime }(x) \ge c \, \frac{-U^{\prime \prime }(x)}{x} > 0 \quad \text{ for } \text{ all } x\ge x_2. \end{aligned}$$ -
(b)
The implication \(\mathrm{(ii)}\, \Rightarrow \,\mathrm{(i)}\) is proved in Menegatti (2014), Proposition 2 (a). To show the opposite implication, we use a similar argument as in the previous proof. First of all we exclude the case where \(\mathrm{ART}(U) < 0\), since by part (a) of this theorem, \(U^{\prime \prime }(x) < 0\) for \(x\ge x_1\). Moreover, from Lemma 2, we know that \(xU^{\prime \prime \prime }(x)\) and \(U^{\prime \prime }(x)\) tend to 0 as \(x \rightarrow +\infty\), hence \(\frac{-xU^{\prime \prime \prime }(x)}{U^{\prime \prime }(x)}\) is an indeterminate form at \(+\infty\). By virtue of L’Hôpital’s rule, it has a finite limit if and only if \(\mathrm{ART}(U)\) is finite and, in this case,
$$\begin{aligned} 0 \le \lim _{x \rightarrow +\infty } \frac{-xU^{\prime \prime \prime }(x)}{U^{\prime \prime }(x)} = -1 + \mathrm{ART}(U) \end{aligned}$$This implies that \(\mathrm{ART}(U)\) is strictly positive (possibly infinite) and the claim follows as above.
-
(c)
The implication \(\mathrm{(ii)}\,\Rightarrow \,\mathrm{(i)}\) follows from Menegatti (2015), Proposition 3. To prove that \(\mathrm{(i)}\, \Rightarrow \, \mathrm{(ii)}\), we proceed as above to show that \(\mathrm{ARRA}_n(U) \ge 0\). Lemma 3 implies that the ratio \(\frac{-xU^{(n)}(x)}{U^{(n-1)}(x)}\) is an indeterminate form at \(+\infty\) and by L’Hôpital’s rule, the limit exists and is finite if and only if \(\mathrm{ARRA}_n(U)\) is finite. In this case,
$$\begin{aligned} 0 \le \lim _{x \rightarrow +\infty } \frac{-xU^{(n)}(x)}{U^{(n-1)}(x)} = -1 + \mathrm{ARRA}_n(U); \end{aligned}$$hence \(\mathrm{ARRA}_n(U)\) is strictly positive (possibly infinite). The result follows as above.
\(\square\)
Proof of Theorem 2
-
(a)
Menegatti (2014), Proposition 1 (b) implies that \(U^{\prime \prime }(x) < 0\) for all x greater than \(x_0\). Theorem 1(b) implies that \(U^{(4)} < 0\) for all x greater than some \(x^*\). Exploiting Lemma 2, we can show, as in the proof of Theorem 1, that
$$\begin{aligned} 0 \le \lim _{x\rightarrow +\infty } -\frac{xU^{\prime \prime \prime }(x)}{U^{\prime \prime }(x)} = -1 +\mathrm{ART}(U). \end{aligned}$$Since \(\mathrm{ART} (U)\) is either strictly greater or smaller than 3, the claim follows.
Assume now that \(\mathrm{ART}(U) = 3\) and \(-\frac{xU^{(4)}(x)}{U^{\prime \prime \prime }(x)}< 3\). The function \(g(x) = -xU^{\prime \prime \prime }(x) - 2U^{\prime \prime }(x)\) is definitely strictly decreasing since its derivative \(g'(x) = -x U^{(4)}(x) - 3U^{\prime \prime \prime }(x)\) is by assumption definitely strictly negative. Moreover, by Lemma 2, it tends to 0 as x goes to \(+\infty\). This implies that \(g(x) > 0\), which is equivalent to the inequality \(-\frac{x U^{\prime \prime \prime }(x)}{U^{\prime \prime }(x)} < 2\), since \(U^{\prime \prime }(x) < 0\). The other case (\(-\frac{xU^{(4)}(x)}{U^{\prime \prime \prime }(x)}>3 \Rightarrow -\frac{xU^{\prime \prime \prime }(x)}{U^{\prime \prime }(x)}> 2\)) can be similarly proved.
-
(b)
Menegatti (2015), Proposition 3 implies that \((-1)^{n-2} U^{(n-1)}(x) > 0\) for all x greater than \(x_0\), while Theorem 1(c) implies that \((-1)^n U^{(n+1)} > 0\) for all x greater than some \(x^*\). Using Lemma 3 and L’Hôpital’s rule, we obtain
$$\begin{aligned} 0 \le \lim _{x\rightarrow +\infty } -\frac{xU^{(n)} (x)}{U^{(n-1)}(x)} = -1 +\mathrm{ARRA}_n(U). \end{aligned}$$Since \(\mathrm{ARRA}_n (U)\) is either strictly greater or smaller than n, the claim follows.
Let now \(\mathrm{ARRA}_n(U) = n\) and \(-\frac{xU^{(n+1)}(x)}{U^{(n)}(x)}< n\) and consider the function \(g(x) = -xU^{(n)}(x) - (n-1) U^{(n-1)}(x)\). We have that \(\lim _{x\rightarrow +\infty } g(x) = 0\) by Lemma 3. Moreover, the derivative \(g'(x) = -x U^{(n+1)}(x) - nU^{(n)}(x)\) is by assumption definitely strictly negative if n is odd (\(U^{(n)}(x) > 0\)) and definitely strictly positive if n is even (\(U^{(n)}(x) < 0\)). Hence g(x) is definitely decreasing if n is odd and definitely increasing if n is even, and, as a consequence, \(g(x) > 0\) for n odd and \(g(x) > 0\) for n even. In both cases, this condition is equivalent to the inequality \(-\frac{x U^{(n)}(x)}{U^{(n-1)} (x)} < n-1\), since \((-1)^{n-2} U^{(n-1)}(x) > 0\). The other case follows in a similar way.
\(\square\)
Proof of Corollary 2
Menegatti (2015), Proposition 3 implies that \((-1)^{j-1} U^{(j)}(x) > 0\) for all x greater than \(x_0\), \(2\le j \le n\). Lemma 3 implies that the ratio \(\frac{-xU^{(j+1)}(x)}{U^{(j)}(x)}\) is an indeterminate form at \(+\infty\) for all \(2\le j\le n\). Then, a recursive application of L’Hôpital’s rule shows that \(\mathrm{ARRA}_j(U)\) exists. Moreover, it is different from j if and only if \(\mathrm{ARRA}_n(U) \not = n\), and it is equal to j otherwise. A recursive application of Theorem 2 yields the result. \(\square\)
Rights and permissions
About this article
Cite this article
De Donno, M., Menegatti, M. Some conditions for the equivalence between risk aversion, prudence and temperance. Theory Decis 89, 39–60 (2020). https://doi.org/10.1007/s11238-020-09745-5
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11238-020-09745-5