Laura Sauci’s research while affiliated with National University of La Rioja and other places

This paper examines trends and persistence in the Greenland ice sheet mass by applying fractional integration methods to a dataset constructed by Mankoff et al. (2020) on ice discharge for seven different regions of Greenland. The adopted empirical framework encompasses a wide range of stochastic processes and is informative about their dynamic and long-run properties. The main finding is that significant changes have occurred in the behaviour of the series of interest in recent years; more specifically, although a deterministic trend is not present, ice discharge in the various regions of Greenland has become a non-stationary, explosive process, with shocks having permanent effects. It appears that, as a result of global warming, the ice mass loss in Greenland has already reached a tipping point and become an irreversible process. JEL Classification: C22; Q54

This paper investigates the time series properties of the temperature and precipitation anomalies in the contiguous USA by using fractional differentiation. This methodology allows to capture time trend components along with properties such as long-range dependence and the degree of persistence. For aggregated data, we find out that long memory is present in both precipitation and temperature since the integration order is significantly positive in the two cases. The time trend is also positive, being higher for the temperature. In addition, observing disaggregated data by states, for the temperature, there are only seven states where the time trend is not significant, with most of them located in Southeast areas, while for the rest of cases, the time trend is significantly positive. All cases exhibit long-range dependence, though the differencing parameter substantially changes from one state to another, ranging from 0.09 in Nebraska and Kansas to 0.18 in Florida and Michigan. For precipitation, the time trend is insignificant in a large number of cases, and the integration order is smaller than for the temperature. In fact, short memory cannot be rejected in fourteen states, and the highest orders of differencing are obtained in Arizona (d = 0.11) and Texas (0.12). In general, we highlight that one cannot draw conclusions about persistence and trends in these two climate-related variables based on aggregate information of the overall USA, given widespread heterogeneity across the states. Tentatively, the degree of dependence across the states seems to be negatively correlated with their level of climate-related risks and the associated preparedness in terms of handling climate change, but this conclusion requires more elaborate research in the future.

In this paper we examine monthly mean temperatures in 40 selected stations in Siberia for the time period January 1937–December 2020 using long range dependence techniques. In particular, we use a fractionally integrated model that incorporates a linear time trend along with a seasonal structure. Our results show first that long memory is present in all stations with significantly positive values for the differencing parameter, though, at the same time the seasonal component seems to be important in all cases. Performing seasonal unit root tests, the results support nonstationary seasonality and working with the seasonal differenced data, the results differ depending on the structure of the error term: if the errors are uncorrelated, long memory is present; however, allowing autocorrelation, this feature disappears in favor of a short memory pattern.

This paper analyses US sea level data using long memory and fractional integration methods. Specifically, monthly data for 41 US stations covering the period from January 1950 to December 2018 are examined. Fractional integration methods suggest that all series exhibit orders of integration in the interval (0, 1), which implies long-range dependence with positive values of the differencing parameter; further, significant positive time trends are found in the case of 29 stations located on the East Coast and the Gulf of Mexico, and negative ones in the case of four stations on the North West Coast, but none for the remaining 8 on the West Coast. The highest degree of persistence is found for the West Coast stations and the lowest for the East Coast ones. Thus, in the event of shocks, more decisive action is required in the case of West Coast stations for the series to revert to their original trend.

In this paper we aim to analyse the degree of mean reversion of public expenditure, revenue and the difference between them, over GDP for 27 European Union (EU) countries. To gain flexibility in the analysis of the order of integration of the variables we apply fractional integration techniques. In general, we find evidence of mean reversion in all countries.

This paper deals with the analysis of the temperatures in a group of 29 stations located in twelve European countries by looking at the coefficients in a linear time trend regression model and allowing for long memory patterns in the error term. The results indicate that long memory is present in practically all cases, and the time trend coefficients are statistically significant in the majority of the cases implying evidence of increasing warming trends. This pattern is particularly noticeable in the case of several stations located across Italy and France, which might be related with micro climates affecting these regions.

This paper investigates the time trend coefficients in the temperatures in 48 US states using monthly data from January 1895 to December 2017, as well as in their anomalies with respect to the base period 1901–2000. For this purpose, we use techniques based on fractional integration, which is a more general approach than the standard methods used in the literature based on integer differentiation. The results indicate that with the exception of 10 states, in the remaining 38, the temperature anomalies have increased across time, with the increase being higher than the one expected under the other more standard approaches. The highest increases correspond to New Jersey and Rhode Island, with an increase of approximately 2.9°C over the last 100 years.

This article examines the long-run Purchasing Power Parity (PPP) hypothesis for 12 Latin American Real Effective Exchange Rates (REERs) using fractional integration techniques. The empirical results, applying parametric approaches, provide evidence of mean reversion in the REERs in the cases of Nicaragua, Belize, Costa Rica, Guyana and Paraguay and lack of it for the remaining seven countries. Employing semiparametric methods, the evidence of mean reversion covers the following countries: Belize, the Dominican Republic, Ecuador and Mexico. Thus, only for Belize and Guyana do we obtain consistent evidence of mean reversion in the real exchange rates. At the other extreme, lack of mean reversion, and thus, lack of PPP, is obtained with both methods in Bolivia, Brazil, Colombia and Venezuela. For the remaining six countries, the results are ambiguous. The results for the PPP theory in Belize and Guyana may show the importance of promoting policies based on exchange rate flexibility and economic liberalization to reach a long-run stability scenario that leads to greater international competitiveness and lower external vulnerability.