سیاست های پولی و خرسندی : ترجیحات بیش از تورم و بیکاری در امریکا لاتین

Using subjective well-being survey data for Latin America, we present evidence that both inflation and unemployment reduce well-being; where the cost of inflation in terms of unemployment, hence the relative size of the weights in a social well-being functions, is about one to eight, almost double of that found for OECD countries. The trade-off, and therefore the misery index, differs across subgroups. For example, the young and left-leaning citizens are more concerned with unemployment than inflation.

Inflation and unemployment are perhaps, for policy makers, the two most important macroeconomic policy targets and of direct concern for ordinary citizens. As any student of monetary policy knows, the general policy problem is cast as one of minimising a social welfare loss function, defined over inflation and unemployment space, subject to a short-term Phillips relation. In a dynamic setting, often complemented with estimations of the sacrifice ratio, the policy problem is to determine the optimal path of disinflation policy. In practice, in policy making circles, the empirical estimation of a welfare loss function is dropped. In fact a pure inflation targeting monetary regime has an implicit loss function that only includes inflation. Since the adoption of inflation targeting by Chile in 1991, other Latin American countries, partly due to its promotion by the IMF, have adopted some variant of an inflation targeting monetary regime. A recent study of seven Latin American countries concluded, “… for most countries, the interest rate setting … does not take into account exchange rate changes or the output gap.” ( Carvalho and Moura, 2008). Given the Okun relation between output and unemployment, they presumably do not take into account the unemployment rate. This policy stance contrasts with OECD countries’ central banks for which both unemployment and inflation are, in principle, concerns of monetary policy. Empirical estimates of social welfare functions have been traditionally plagued with problems associated with a priori imposition of preferences structure (see Woodford, 2001). However, recently these problems have been avoided by directly estimating the inflation and unemployment weights in the loss function from happiness surveys. The literature using this data under the broad concept of well-being uses interchangeably happiness and life satisfaction. Most studies on unemployment–inflation–happiness use life satisfaction. While the concepts are somewhat different, responses in different surveys are highly correlated. Also, a number of validation studies suggest that happiness questions reveal something meaningful regarding well-being. Self-reported happiness has high test–retest correlations and correlates well with variables used in psychometric analysis. For example, happiness scores correlate well with the demographic characteristics of the respondents across countries, which would not be so if the data was just noise.2 Thus the inflation–unemployment–happiness research assumes that self-reported happiness scores are a measure of true utility with a high signal to noise ratio. However, most studies on happiness draw almost exclusively on data from OECD countries and as far as we know there is no inflation–unemployment–happiness study on Latin America. A priori we expect different estimations for Latin American countries relative to OECD countries as Latin American countries are subject to more shocks, have higher macroeconomic volatility and have little to no social safety nets – particularly unemployment insurance schemes – relative to OECD countries. The objective of this paper is to fill this lacuna and present estimations of the weights of inflation and unemployment in the loss function of Latin American countries. The estimated weights are used to draw out the ramifications for monetary policy particularly by contrasting citizen's preferences with those of inflation targeting central bankers. They are also used to contrast with the weights estimated for OECD countries. The paper is structured as follows. First, we briefly review the literature on inflation–unemployment–happiness that uses country panel data. Second, we present our research strategy discussing the data and methodology used. Third, we present our estimations of the unemployment–inflation trade-off for the general population and for sub-groups. We end the article with a discussion of the policy ramifications of our findings

Macroeconomists and policy makers assume that unemployment and inflation are two “bads”, that is, they reduce social welfare. The policy-relevant question is if there is a trade-off between them then, what is the marginal rate of substitution hence the optimal disinflation policy? In this paper we present direct evidence that unemployment and inflation reduce happiness. The evidence shows, however, that unemployment causes more unhappiness than inflation. Unlike OECD countries where the unemployment to inflation trade-off, i.e. the cost of inflation in terms of unemployment for a constant level of happiness, is about four, in Latin America it is about eight. Further, the trade-off differs substantially across subgroups. For the young (aged 18–24 years) it is higher at 12. There is also evidence that the political leaning of a person effects the size of the trade off; left leaning citizens are more concerned with unemployment, with a trade off of 10, while right leaning citizens do not care about unemployment only inflation. Thus right-wing partisans are in line with those monetary professionals who advocate pure inflation targeting rule for central banks. However, such advocacy is increasingly divorced from the opinions of LAC citizens who are increasingly left leaning and with the youth who, given the population pyramid, will also increase as a proportion of the population. Using the estimated trade off the misery index reveals that the weighted misery index differs in level (it is higher) and change (with an increase rather than a fall) from the commonly used non-weighted index and from that of the pure inflation targeters. Thus the latter two give false signals. Therefore, should central bankers target happiness instead of inflation? Perhaps not, the evidence presented in this paper, combined with the low frequency of happiness data, may not be sufficient but is in line with the conclusion of Di Tella and MacCulloch (2007): “Against this backing, we argue that happiness data can be quite useful to central banks”. The fact that unemployment is more costly than inflation is particularly relevant in discussions of the desirability of gradualist relative to draconian monetary policy options, a choice more often facing Latin American central banks relative to OECD countries’ central banks.