Skip to main content

Aversion to Risk

  • Chapter
  • First Online:
Illustrating Finance Policy with Mathematica

Abstract

Aversion to risk is a central feature of decisions. This chapter introduces the mathematical interpretation of aversion to financial risk and insurance.

This is a preview of subscription content, log in via an institution to check access.

Access this chapter

Chapter
USD 29.95
Price excludes VAT (USA)
  • Available as PDF
  • Read on any device
  • Instant download
  • Own it forever
eBook
USD 74.99
Price excludes VAT (USA)
  • Available as EPUB and PDF
  • Read on any device
  • Instant download
  • Own it forever
Softcover Book
USD 99.00
Price excludes VAT (USA)
  • Compact, lightweight edition
  • Dispatched in 3 to 5 business days
  • Free shipping worldwide - see info
Hardcover Book
USD 99.00
Price excludes VAT (USA)
  • Durable hardcover edition
  • Dispatched in 3 to 5 business days
  • Free shipping worldwide - see info

Tax calculation will be finalised at checkout

Purchases are for personal use only

Institutional subscriptions

Notes

  1. 1.

    In mathematical notation, we seek the greatest x for which

    CDF(BinomialDistribution(100, .5), x)≤.01

    Because Mathematica cannot solve this, create a loop that obtains the answer by decrementing a variable (x) from the maximum number of risks (and where the CDF is 1) until the value of the CDF falls below the target (here .01):

    x=100 ;

    While[CDF[BinomialDistribution[100,.5],x]>.01,x=x-1]

    The outcome of the loop is that x is the largest number of wins where the cumulative probability is at or under 1%. To decrement x , instead of x=x-1 we can write x-- .

  2. 2.

    R. Mehra & E.C. Prescott, The Equity Premium: A Puzzle, 15 J. Monetary Econ. 145, 154 (1985) (collecting estimates from various scholars mostly from 1 to 2 and positing that its value should be under ten). Others, focusing on the pricing of agricultural commodities produce extremely low estimates of risk aversion; see generally Rob Raskin & Mark J. Cochran, Interpretations and Transformations of Scale for the Pratt-Arrow Absolute Risk Aversion Coefficient: Implications for Generalized Stochastic Dominance, 11 W.J. Agric. Econ. 204, 205 & Table 1 (1986) (collecting estimates, all near zero).

  3. 3.

    Karel Janecek, What Is a Realistic Aversion to Risk for Real-World Individual Investors? (2004), www.sba21.com/personal/RiskAversion.pdf [perma.cc/S86X-YD8N]. See also S. Kandel & R.F. Stambaugh, Asset Returns and Intertemporal Preference, 27 J. Monetary Econ. 39 (1991) (also allowing the plausibility of a value of 30 for the coefficient of risk aversion).

  4. 4.

    We can use Mathematica to solve this differential equation. The result is \(u(w) = \frac{{w^{1 - a} {C}_{1} }}{1 - a} + {C}_{2}\). The differential equation approach gave us a function of wealth w and risk aversion, a, but two additional parameters appeared, C1 and C2. Those are constants of integration. Solving differential equations without having specified where to place the resulting curves on the plane leaves constants of integration. Thinking back to the curve of Fig. 1, changing C2 moves that curve in the vertical direction and changing C1 moves it in the horizontal direction. If we had a specific value of utility that some amount of wealth should create, then the constants of integration would make the resulting function comply with this specification. However, because utility of wealth is not measurable, we do not have any such specification. Rather, the convention is to generalize away the constants of integration. They disappear by taking neutral values, zero for C2 and one for C1.

  5. 5.

    The raw solution is \(u(w) = {C}_{1} {\log}\left( w \right) + {C}_{2}\), where \({\log}\left( \, \right)\) is the natural logarithm, often noted as \(ln\). Replacing C1 with 1 and C2 with 0 gives the clean result in the text.

  6. 6.

    A fuller analysis appears in my article Nicholas L. Georgakopoulos, Predictability and Legal Evolution, 17 Int’l Rev. L. & Econ. 475–89 (1997).

Author information

Authors and Affiliations

Authors

Corresponding author

Correspondence to Nicholas L. Georgakopoulos .

Appendix: Code for Figure 2

Appendix: Code for Figure 2

(*loop of insolvency values of the insurer*) Clear[x, y] y = 1; array1 = Table[y = y/2, {150}]; array2 = 1 - array1; loopsol =    Flatten@{Reverse@array1, Drop[array2, 1], 1(*drop the      duplicate 1/2, add a final 1*)}; (*loop of values of risk aversion*) Clear[x, y, ex] ex = 4; looprac =   Flatten@{0,     Table[x^ex , {x, .2^(1/ex), 32^(1/ex), (32^(1/ex) - .25^(1/ex))/      15}]}; mintarg = 0;(*solvency loop minimum*) maxtarg = .9999999999;(*solvency loop maximum*) ni = 100;(*number of insured*) ps = .5;(*probability of success of each insured, i.e., probability of the gamble winning, of not suffering harm but having gain*) (*set premium so that the insurer becomes insolvent with only target (say 1%) probability: for how many wins is the CDF[Binomial[ ],wins] equal to target?*) target =.; cra =.; w0 = 3;(*initial wealth of each insured*) wg = 1;(*size of wager of each insured in the double or nothing risk/venture*) ci = .25;(*cost of insolvency, fraction lost to lawyers and auctioneers*) surfc = Table[{  x = ni;  While[CDF[BinomialDistribution[ni, ps], x] > target, x–];  pm = (wg (ni - x))/ni;(*premium payment by each insured*)  (*Print[“The probability of failure is under ”,target 100,”% at ”,  x,” wins, at most. The payment is ”,pm//N,”.”]*)  wt[target] = x;(*store the breakeven wins*)  x =.;  mibi = ni - (ni pm)/wg;(*equal to wt[target];   number of wins making the mutual insurance barely insolvent,   result of Solve[wg (ni-x)\[Equal]ni pm,x], i.e.,     how many wins make the lost wagers equal to the assets of the insurer?*)    (*if fewer than mibi wins,     then the losers suffer uncompensated harm.     In these cases we need the amount of the uncompensated harm. The assets of the mutual insurer, ni pm, reduced by the cost of insolvency ci to (1-ci)ni pm, are divided proportionately among the number of actual non-wins, ni-wins; therefore instead of losing wg they lose wd minus the distributed assets.*)    uh = wg - ((1 - ci) ni pm)/(ni - wins);(*uncompensatd harm. the lost wager reduced by the distributed assets, which are the assets of the insurer ni pm divided by the losses, ni-wins; Verification: should be zero when wins= mibi except in the case of no insurance, target=1, when it will be 1.*)    lii = w0 - uh - pm;(*wealth of losers if the insurer becomes insolvent*)    (*the rest of the time, with probability wins/ni,     the insured wins and ends up with wealth w0 + wg-pm*)    wii = w0 + wg - pm;(*winning insured wealth if the insurer becomes insolvent*)    wmni = (ni pm - (ni - wins) wg)/ni;(*per insured wealth of mutual insurer if it does not become insolvent, after if compensates all harmed insureds, distributed to the ni members. Becomes zero if the number of wins is mibi*)    wis = w0 + wg - pm + wmni;(*wealth of winners if the insurer stays solvent, i.e., for wins>mibi*)    lis = w0 - pm + wmni;(*wealth of losers if the insurer stays solvent, i.e., for wins>mibi; note: the insurance is full, compensating their entire loss, else we would need to subtract a deductible*)    eu = Sum[      PDF[BinomialDistribution[ni, ps], wins] If[        wins <= wt[target](*mibi*),        (ni - wins)/ni u[lii, cra]  +         wins/ni u[wii, cra], (ni - wins)/ni u[lis, cra] +          wins/ni u[wis, cra]],      {wins, 0, ni}];(*expected utility: probabilty of being a loser or a winner, times the resulting utility of wealth*)    (*corrsponding certainty-equivalent wealth*)    Off[Solve::ifun];    ceiA =      x /. (Solve[u1[x, a] == eu, x] //          Simplify)[[1]];(*certainty equivalent*)    cei1 =      x /. (Solve[u2[x] == eu, x] //          Simplify)[[1]];(*certainty equivalent if a=1*)    twi = If[cra == 1, cei1,       ceiA /. a -> cra];(*get certainty equivalent*)    (* get certainty equivalent with no insurance:*)    cea = x /. (Solve[          u1[x, a] == (1 - ps) u1[w0 - wg, a]  +  ps u1[w0  +  wg, a], x] //          Simplify)[[1]];    ce1 = x /. (Solve[u2[x] == (1 - ps) u2[w0 - wg] + ps u2[w0 + wg],           x] // Simplify)[[1]];    On[Solve::ifun];    tw = If[cra == 1, ce1,       cea /. a -> cra];(*get certainty equivalent*)    cra, target, twi - tw},   {cra, looprac}, {target, loopsol}]; mincex = 0; maxcex = 32; mincey = 0; maxcey = maxtarg; mincez = 0; \ maxcez = .55; inswelfeff1 = Graphics3D[{Black,    Line /@ surfc,    Line /@ Transpose@surfc, (*labels*)    Text[“Risk Av’n Coeff’t”, {.82 maxcex, mincey, mincez  +  .025}],    Text[“Insolv’cy %”, {mincex, .8 maxtarg, mincez + .03}],    Text[“Gain in Cert. Equiv’t”, {0, -.03, .75 maxcez}],    Thickness[.002],(*axes*)    Line[{{mincex, mincey, mincez}, {maxcex, mincey, mincez}}],    Line[{{mincex, mincey, mincez}, {mincex, maxcey, mincez}}],    Line[{{mincex, mincey, mincez}, {mincex, mincey, maxcez}}],    Thickness[.004],    Line[surfc[[1]]],    Line[surfc[[7]]],    Line[surfc[[-1]]]    },   BoxRatios -> {1, 1, .6}, Boxed -> False,   ImageSize -> 7 imgsz,   BaseStyle -> fntandsz,   ViewPoint -> {1.75, -3, .65}]

Rights and permissions

Reprints and permissions

Copyright information

© 2018 The Author(s)

About this chapter

Check for updates. Verify currency and authenticity via CrossMark

Cite this chapter

Georgakopoulos, N.L. (2018). Aversion to Risk. In: Illustrating Finance Policy with Mathematica. Quantitative Perspectives on Behavioral Economics and Finance. Palgrave Macmillan, Cham. https://doi.org/10.1007/978-3-319-95372-4_10

Download citation

  • DOI: https://doi.org/10.1007/978-3-319-95372-4_10

  • Published:

  • Publisher Name: Palgrave Macmillan, Cham

  • Print ISBN: 978-3-319-95371-7

  • Online ISBN: 978-3-319-95372-4

  • eBook Packages: Economics and FinanceEconomics and Finance (R0)

Publish with us

Policies and ethics