Pamela Giustinelli, "Non-Parametric Bounds on Quantiles under Monotonicity Assumptions, With an Application to the Italian Education Returns", Journal of Applied Econometrics, Vol. 26, No. 5, 2011, pp. 783-824. Binary data files are zipped in the file pg-binary.zip. ASCII data files are zipped in the file pg-ascii.zip. Matlab code files are zipped in the file pg-matlab-codes.zip. Unix/Linux users should use "unzip -a" when unzipping the latter two files. The data come from the Survey on Household Income and Wealth conducted by the Bank of Italy and are publicly available at: http://www.bancaditalia.it/statistiche/indcamp/bilfait The file ``dati_pamela.dta" was obtained by: 1) pooling together the 1993, 1995, 1998, 2000 and 2002 waves; 2) selecting only males between 30 and 55, employed full-time, not in the agricultural sector; 3) allocating to one interview wave only those individuals that were interviewed more than once (i.e. those belonging to the panel component of SHIW). Notice: the file ``dati_pamela.dta", obtained after the above manipulations, was available to me by previous works on the same data, and contains a number of variables used in those works, but not in the current paper. Hence, I provide this file as the starting data set for application in the current paper. This file was subsequently modified as described below. The file "dati_qboundsedu.dta" contains manipulations of some of the variables in "dati_pamela.dta" (through the Stata code "BoundsEduc.do"), involving creation of the deflator variable, re-codification of the education variables for use in the empirical application and drop of variables not used in the application itself. The data GNP deflator for various years is gathered in the file "deflattori.xls", obtained from ISTAT (http://www.istat.it/). There is also an ASCII version in deflattori.csv. *** List and description of variables in "dati_qboundsedu": (1) dlnyora: This variable is obtained by appling the appropriate deflator to lnyora, which in turn was obtained by applying the natural logaritm to a hourly net wage variable calculated as: (net annual income)/(months worked * hours worked per week * 4), and where the average number of hours worked per week includes overtime. Annual income is expressed in Euros and is net of taxes and social security contributions. (2) istru4: This variable is obtained by re-coding istru (years of education corresponding to the maximum qualification obtained by the individual) into the following 4 categories: a) Elementary or less - This category groups people with either an elementary graduation degree or no educational quali cation, normally corresponding to 5 or less years of formal education. In fact, since the SHIW questionnaire asks individuals to report their highest attained qualification, respondents recorded as having an elementary qualification may actually have more than 5 years of education while never obtaining a junior high school degree. b) Junior high degree - At least 8 years. c) High school diploma - At least 3-to-5 years of schooling after junior high, depending on the attended type. d) College degree or higher - At least 3 additional years of education after any high school program of 5 years. This group includes also individuals with Masters and Ph.D.s (3) istruMHS: This variable is obtained by re-coding istrum (years of education corresponding to the maximum qualification obtained by the mother of the individual) into the following 2 categories: a) Junior high or less b) High school or more