Fairly similar rules in Peru's 1979 and 1993 constitutions have governed the relation between the executive and the legislature in the country during the last three decades, but throughout these years the balance of power between the executive and Congress has been very unstable. During President Alan García's first term in office (1985–90) we witnessed a relatively strong Congress limiting the power of a powerful executive. From 1992 to 2000, a weak Congress subordinated itself to Alberto Fujimori and his military allies. Finally, during Alejandro Toledo's government (2001–6) we find a stronger Congress, although less so than in the 1980s. How can we explain this stark variation?

Levitt focuses on two variables to answer this question: party organisational strength and the elite's adherence to constitutional norms. First, he suggests, strong political parties (as opposed to weakly organised political movements) empower the legislature by using constitutional rules to check the executive. It is not only the strength of opposition parties that causes this enhancement of congressional power: competition within the incumbent party also contributes to strengthening the legislature. Second, political elites that are loyal to constitutional norms will respect (to a greater extent) the rules of the game, checking more thoroughly the actions of the executive.

According to Levitt, these two variables explain the outcomes discussed above. Although feeble in a comparative perspective, during the 1980s Peruvian political parties were fairly well organised and political elites adhered to the Constitution. Thus, Congress became a significant institutional actor during García's term, with even the Alianza Popular Revolucionaria Americana (American Popular Revolutionary Alliance, APRA), the incumbent party, checking the power of the president. The severe economic and political crises of the 1980s weakened parties’ organisation and political elites’ adherence to constitutional rules, leading to an enhancement of executive power during Fujimorismo. Finally, when Fujimori fell due to a corruption scandal, old and new parties regained some of their strength and became more respectful of constitutional norms. Executive power declined, although lower scores in both variables explain the reduced power of the legislature when compared with the 1980s. The final chapter aims to test these findings in Latin America.

The book highlights two important aspects of current Latin American politics. First, it shows the limits of formal institutionalism as a means of analysing politics in the region. Formal institutions have weak leverage when explaining presidential or congressional effective power. Second, the study shows that generalisations about institutional weakness in the region are also too simplistic. The variation found from episode to episode shows that legislatures can be meaningful institutions, even in a country with such a weak legislative tradition as Peru.

Although it provides an interesting description of the recent Peruvian political process, the book has two weaknesses: part of its argument and its theoretical framework. With regard to the argument, the first variable – party organisational strength – certainly seems quite relevant to explaining why legislatures become active actors in politics. Strong parties will use the institutional weapons included in the Constitution to check and balance the executive. Nonetheless, this first factor needs a more thorough discussion of the independent effect of political pluralism for this outcome. The author believes that the size of the president's coalition can only explain ‘so much’ variation in the power of the executive (p. 6), and he is partially right. As the APRA case (1985–90) shows, an incumbent party with a majority in Congress became a limit to the president's power. But this case should not make the author disregard the importance of political pluralism as an independent factor restraining concentration of power in the executive. APRA's majority was slight, and the other parties in Congress were relatively competitive. Some degree of political pluralism seems necessary for limiting executive power.

However, it is the second variable – elite adherence to the Constitution – that presents the main problem. Can political elites lose and regain constitutional values so quickly as to explain abrupt changes like the ones described? It seems unlikely that changes in elite values can operate so swiftly. Unfortunately, the author never deals properly with this interesting idea; there is no systematic effort in the book to measure the change in this variable through time. In the chapter Levitt devotes to constitutional adherence (chapter 3) we find a long description of political events in Peru and just in passing some mention of alleged changes in elites’ constitutional values. But there is no systematic analysis of how they may have operated to undermine congressional authority.

If Levitt wanted to explain such abrupt changes through this value-centred argument, perhaps he could have posited that party elites are likely to be more respectful of constitutional norms than independent politicians. It seems plausible to argue that the socialisation of party elites in electoral politics may make them more concerned with constitutional norms, thus helping to explain the variation in the Peruvian case. But if that is the case, then the first factor – party strength – will be doing most of the explanatory work, as strong parties will be conducive to party elites in power. In any case, much more should have been said about the precise meaning and effect of elites’ constitutional adherence.

Arguably the work's main weakness is that it does not add much to extant theoretical frameworks. We already know that formal institutionalism by itself has clear limits explaining politics in the region and, more broadly, the developing world. We also know that horizontal institutions in Latin America are not doomed to be weak, especially after recently witnessing legislatures deposing presidents and courts ruling against executives. And the fact that Congress is expected to be stronger in a competitive and organised party system does not seem a novel finding either. Sadly, then, the potentially most interesting contribution – the importance of norms as a determinant of political conduct – is the one that remains least developed.