The limit of law: factors influencing the decision to make harmful acts illegal


This paper examines factors affecting the decision of whether or not to make certain harmful acts illegal. It considers factors related both to the cost of law enforcement and to the crime commission decision. On the enforcement side, illegality is limited by the existence of fixed notice and response costs, which are unrelated to the harm from the act, and also by costs of imposing punishment. In addition, illegality is limited by a finite marginal productivity of detection, one cause of which is legal error. On the commission side, illegality is limited if offenders have strictly positive benefits from committing the act. The paper concludes by examining how the optimal scope of law is affected by its “expressive function”, the idea that some people are deterred by the mere fact that an act is illegal. We specifically ask how the scope of law changes if more people behave in this way. The answer depends on whether “efficient” violations of the law are possible, which in turn depends on whether offenders’ gains are counted in welfare.

This is a preview of subscription content, access via your institution.


  1. 1.

    Advances in GPS technology, however, have greatly reduced this cost, thus making pure Pigovian taxation a more viable enforcement option.

  2. 2.

    See, for example, McAdams (2017), Dharmapala and Garoupa (2003), Sunstein (1996), Cooter (1998), and Kahan (1998).

  3. 3.

    The debate began with Stigler (1970). See Lewin and Trumbull (1990), Friedman (2000, p. 230), and Curry and Doyle (2016) for fuller discussions.

  4. 4.

    Following this literature, throughout the article we refer to the government’s choice as criminalization, although the model is equally applicable to the illegalization of acts which are regulated through public and non-criminal enforcement.

  5. 5.

    Thus, for example, if s is a fine, \(\sigma =-1\), so the overall social cost of imposing s, including the cost to the offender, would be \((1+\sigma )s=0\), because the cost to the offender is exactly offset by the revenue received by the government. However, if \(\sigma >-1\), there is a net social cost of imposing s.

  6. 6.

    In truth, notification costs will not be completely fixed. For example, the government will choose how much effort to devote to notifying the public about a law based on the perceived harm. Think, for example, of public service announcements and billboards about drunk driving laws, or the decision of how far apart to space speed limit signs. We can safely abstract from this consideration because some notice costs are definitely fixed, and any variable costs can be captured by our other cost categories.

  7. 7.

    For purposes of isolating factors limiting optimal criminalization, we focus on the most studied case of risk-neutrality. When this assumption is relaxed, gains can emerge from the deliberate random enforcement of criminal laws.

  8. 8.

    There are models in which the optimal sanction is not maximal, even if it is a fine; for example, when offenders can invest in efforts to avoid capture (Malik 1990), when juries resist convicting offenders if punishment appears too harsh (Andreoni 1991), when type 1 errors (false convictions) are accounted for in welfare (Miceli 1991), and when criminal benefits are state-dependent (Mungan 2010). However, none of these factors is relevant here.

  9. 9.

    We note that wihout the expressive function of the law, it follows that criminalization coupled with symbolic enforcement (e.g. \(s=0\) or \(x=0\)) cannot be optimal. This follows from a simple comparison of (2) and (3). We discuss the case where the law has an expressive function in Sect. 4.

  10. 10.

    We assume throughout our analysis that \(x^{*}(h)\) is a proper maximizer of welfare.

  11. 11.

    This follows the conventional view of the effects of legal error (see, e.g., Png 1986). For a recent debate on the impact of type 1 error on deterrence, see Lando (2006), Garoupa and Rizzolli (2012), Lando and Mungan (2018), and Mungan (2017). If type 1 errors do not reduce deterrence, as considered in Lando (2006), the results presented in proposition 1 may become more relevant.

  12. 12.

    See the references in note 2 above.

  13. 13.

    See, for example, Coşgel and Miceli (2017), and Shavell (2002), who discusses morality and law as alternative regulators of behavior.

  14. 14.

    This formulation treats reflexive obedience to the law as an all or nothing decision—either people have internalized the sentiment or not. An alternative approach would be to assume that people have varying “moral costs” of breaking the law (apart from any threatened legal sanction), which range from zero to infinite. The implications of this alternative approach are pointed out in note 16 below.

  15. 15.

    Stigler (1970) first raised this issue.

  16. 16.

    In the current formulation, we examine this question by doing a comparative static on \({\hat{h}}\) with respect to \(\beta\). In the more general formulation suggested in note 14 above, where people have varying moral costs of breaking the law, an increase in overall morality would be reflected by a rightward shift (in the sense of first-order stochastic dominance) in the distribution of this cost. The qualitative conclusions of such an approach would mirror the conclusions obtained from the simpler formulation in the text.

  17. 17.

    See Miceli (2019, Chapter 8) for a preliminary version of this result.


  1. Andreoni J (1991) Reasonable doubt and the optimal magnitude of fines: should the penalty fit the crime? Rand J Econ 22:385–395

    Article  Google Scholar 

  2. Bowles R, Faure M, Garoupa N (2008) The scope of criminal law and criminal sanctions: an economic view and policy implications. J Law Soc 35:389–416

    Article  Google Scholar 

  3. Becker G (1968) Crime and punishment: an economic approach. J Polit Econ 76:169–217

    Article  Google Scholar 

  4. Cooter R (1998) Expressive law and economics. J Legal Stud 27:585–607

    Article  Google Scholar 

  5. Coşgel M, Miceli T (2017) Toward an economic theory of religious belief and the emergence of law. J Inst Theor Econ 175:692–713

    Article  Google Scholar 

  6. Curry P, Doyle M (2016) Integrating market alternatives into the economic theory of optimal deterrence. Econ Inquiry 54:1873–1883

    Article  Google Scholar 

  7. Dharmapala D, Garoupa N (2003) The condorcet jury theorem and the expressive function of law. Am Law Econ Rev 5:1–31

    Article  Google Scholar 

  8. Friedman D (2000) Law’s order: what economics has to do with law and why it matters. Princeton University Press, Princeton

    Google Scholar 

  9. Garoupa N, Klerman D (2002) Optimal law enforcement with a rent-seeking government. Am Law Econ Rev 4:116–140

    Article  Google Scholar 

  10. Garoupa N, Obidzinski M (2011) The scope of punishment. Eur J Law Econ 31:237–247

    Article  Google Scholar 

  11. Garoupa N, Rizzolli M (2012) Wrongful convictions do lower deterrence. J Inst Theor Econ 168:224–231

    Article  Google Scholar 

  12. Hart H (1982) Punishment and responsibility: essays in philosophy of law. Oxford University Press, Oxford

    Google Scholar 

  13. Holmes O (1897) The path of the law. Harvard Law Rev 10:457–478

    Article  Google Scholar 

  14. Kahan D (1998) Social meaning and the economic analysis of crime. J Legal Stud 27:609–622

    Article  Google Scholar 

  15. Lando H (2006) Does wrongful conviction lower deterrence? J Legal Stud 35:327–337

    Article  Google Scholar 

  16. Lando H, Mungan M (2018) The effect of type-1 error on deterrence. Int Rev Law Econ 53:1–8

    Article  Google Scholar 

  17. Lewin J, Trumbull W (1990) The social value of crime? Int Rev Law Econ 10:271–284

    Article  Google Scholar 

  18. Malik A (1990) Avoidance, screening, and optimum enforcement. Rand J Econ 21:341–353

    Article  Google Scholar 

  19. McAdams (2017) The expressive powers of law: theories and limits. Harvard University Press, Cambridge

  20. Miceli T (1991) Optimal criminal procedure: fairness and deterrence. Int Rev Law Econ 11:3–10

    Article  Google Scholar 

  21. Miceli T (2019) The paradox of punishment: reflections on the economics of criminal justice. Palgrave Macmillan, Cham

    Book  Google Scholar 

  22. Mungan M (2010) A note on the effects of state-dependent benefits on optimal law enforcement. Rev Law Econ 6:97–106

    Article  Google Scholar 

  23. Mungan M (2012) The scope of criminal law. In: Harel A, Hylton K (eds) Research handbook on the economics of criminal law. Edward Elgar, Cheltenham

    Google Scholar 

  24. Mungan M (2017) Wrongful convictions, deterrence, and stigma dilution. Supreme Court Econ Rev 25:199–216

    Article  Google Scholar 

  25. Png I (1986) Optimal subsidies and damages in the presence of judicial error. Int Rev Law Econ 6:101–105

    Article  Google Scholar 

  26. Polinsky AM, Shavell S (2007) The theory of public law enforcement. In: Polinsky AM, Shavell S (eds) Handbook of law and economics. Elsevier, Amsterdam

    Google Scholar 

  27. Shavell S (1985) Criminal law and the optimal use of nonmonetary sanctions as a deterrent. Columbia Law Rev 85:1232–1262

    Article  Google Scholar 

  28. Shavell S (1991) Specific versus general enforcement of law. J Polit Econ 99:1088–1108

    Article  Google Scholar 

  29. Shavell S (2002) Law versus morality as regulators of conduct. Am Law Econ Rev 4:227–257

    Article  Google Scholar 

  30. Stigler G (1970) The optimum enforcement of laws. J Polit Econ 78:526–536

    Article  Google Scholar 

  31. Sunstein C (1996) On the expressive function of law. Univ Pennsylvania Law Rev 144:2021–2053

    Article  Google Scholar 

Download references

Author information



Corresponding author

Correspondence to Thomas J. Miceli.

Additional information

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

We thank two anonymous referees for valuable comments and suggestions.



Proof of Proposition 1

Parts (i)–(iv): First, consider the case where \({\underline{b}}=0\). Note that when \(c=0\), \(\sigma =-1\), and

$$\begin{aligned} \frac{\partial W_{1}}{\partial x}= & {} f(p(x)s)p^{\prime }(x)s[h-p(x)s]-1<0 \text { if} \\ L(x)> & {} h\text { where }L(x)\equiv \frac{1}{f(p(x)s)p^{\prime }(x)s}+p(x)s \end{aligned}$$

L(x) has a strictly positive lower bound, \({\bar{h}}\), when \(p^{\prime }(0)\) is finite. Thus, \(\frac{\partial W_{1}}{\partial x}<0\) for all x and all \(h<{\bar{h}}\). Moreover, \(W_{0}(h)-W_{1}(0,h,\sigma ,k,c)=k\ge 0\), and, thus, \(W_{0}(h)\ge W_{1}(0,h,\sigma ,k,c)\) for all x and all \(h<{\bar{h}}\) when \(c=0\), \(\sigma =-1\), and \(p^{\prime }(0)\) is finite. Additionally, \(\frac{ \partial W_{1}}{\partial c},\frac{\partial W_{1}}{\partial \sigma }<0\) and \(\frac{\partial W_{0}}{\partial c},\frac{\partial W_{0}}{\partial \sigma }=0\) for all x, and, therefore \(W_{0}(h)\ge W_{1}(0,h,\sigma ,k,c)\) for all x and all \(h<{\bar{h}}\) whenever \(p^{\prime }(0)\) is finite.

When \(\underset{x\rightarrow 0}{\lim }p^{\prime }(0)=\infty\), it follows that \(\underset{x\rightarrow 0}{\lim }L(x)=0\) and \(L(x)>0\) for all \(x>0\). Thus, \(\underset{x\rightarrow 0}{\lim }[W_{0}(0)-W_{1}(0,0,\sigma ,k,c)]=k\ge 0\), and, therefore, \(W_{0}(0)>W_{1}(x,0,\sigma ,k,c)\) for all \(x>0\). But, \(\frac{\partial W_{1}}{\partial c},\frac{\partial W_{1}}{\partial \sigma },\frac{\partial W_{1}}{\partial k}<0\), and, thus, \(\underset{ x\rightarrow 0}{\lim }[W_{0}(0)-W_{1}(x,0,\sigma ,k,c)]>0\) and \(W_{0}(0)-W_{1}(x,0,\sigma ,k,c)>0\) for all x whenever \(\sigma >-1\), \(c>0\) or \(k>0\). This implies that whenever one of these conditions are met, there exists \({\bar{h}}>0\) such that \(W_{0}(0)-W_{1}(x,h,\sigma ,k,c)>0\) for all x and all \(h<{\bar{h}}\).

The above observations together imply that if \({\underline{b}}=0\) it follows that whenever \(\sigma >-1\), \(c>0\), \(k>0\) or \(p^{\prime }(0)\) is finite, there exists some \({\hat{h}}>0\) such that \(h<{\hat{h}}\) implies that any policy with \(\psi =1\) and \(x>0\) is inefficient.

Part (v): The derivative of welfare can be expressed as

$$\begin{aligned} \frac{\partial W_{1}}{\partial x}=(h-{\hat{b}})f({\hat{b}})\frac{\partial {\hat{b}} }{\partial x}-1 \end{aligned}$$

We previously noted that if \({\underline{b}}=0\), it is optimal to criminalize all harmful acts, i.e. \({\hat{x}}(h)>0\) for \(h>0\) where \({\hat{x}}\) denotes the optimal solution when \({\underline{b}}=0\), and that \({\hat{x}}\) is increasing in h. Since, \(p(0)=0\) and \(p^{\prime }>0\), then if \({\underline{b}}>0\) there exists \({\underline{x}}>0\) such that \({\hat{b}}={\underline{b}}\) for \(x\le {\underline{x}}\). Thus, \(\frac{\partial {\hat{b}}}{\partial x}=0\) for \(x\le {\underline{x}}\), which implies that (13) negative in this range. Now define \({\hat{h}}\) by \({\underline{x}}={\hat{x}}(\hat{h })\). It follows that \(h<{\hat{h}}\) implies \(x^{*}=0\). \(\square\)

Proof of Proposition 2:

Note that when \(c=0\) and \(\sigma =-1\),

$$\begin{aligned} \frac{\partial W_{1}}{\partial x}= & {} f({\hat{b}}(x))(p^{\prime }(x)-q^{\prime }(x))s[h-{\hat{b}}(x)]-1<0\text { if} \\ {\hat{L}}(x)> & {} h\text { where }{\hat{L}}(x)\equiv \frac{1}{f(p(x)s)(p^{\prime }(x)-q^{\prime }(x))s}+{\hat{b}}(x) \end{aligned}$$

When \(p^{\prime }(0)-q^{\prime }(0)\) is finite, \({\hat{L}}(x)\) has a strictly positive lower bound \({\bar{h}}\). Thus, the proof follows from arguments very similar to those made in the proof of proposition 1. \(\square\)

Proof of Proposition 3

(i) Totally differentiating (10) gives:

$$\begin{aligned}&-\int _{ps}^{{\bar{b}}}(\alpha b-{\hat{h}})f(b)dbd\beta +(1-\beta )\int _{ps}^{ {\bar{b}}}bf(b)dbd\alpha \\&\quad -(1-\beta )[1-F(ps)]d{\hat{h}}=E(b)d\alpha -d{\hat{h}} \end{aligned}$$

From this we compute:

$$\begin{aligned} \frac{\partial {\hat{h}}}{\partial \beta }=\frac{\int _{ps}^{{\bar{b}}}(\alpha b- {\hat{h}})f(b)db}{1-(1-\beta )[1-F(ps)]} \end{aligned}$$

The sign of this expression depends on the numerator, which is (11).

(ii) We also get

$$\begin{aligned} \frac{\partial {\hat{h}}}{\partial \alpha }=\frac{E(b)-\int _{ps}^{{\bar{b}} }bf(b)db}{1-(1-\beta )[1-F(ps)]} \end{aligned}$$

which is positive. Thus, as a larger fraction of offender gains are counted, the scope of law contracts. \(\square\)

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Miceli, T.J., Mungan, M.C. The limit of law: factors influencing the decision to make harmful acts illegal. Econ Gov (2021).

Download citation


  • Crime
  • Illegality
  • Law enforcement
  • Expressive function of law

JEL Classification

  • K14
  • K42