GARCH with generalized Pareto tail

This paper proposes the use of a spliced distribution with generalized Pareto tail for financial risk management. The proposed distribution is tailored to flexibly capture the heavy tail in asset return distribution. The parameters of the distribution can be estimated jointly with a conditional heteroskedasticity model. The estimated parameters can then be used to produce tail risk forecasts for risk management purposes. The use of the proposed distribution is illustrated by evaluating tail risk forecasts for a number of major stock indices.

Paradoxical impact of sprawling intra-Urban Heat Islets: Reducing mean surface temperatures while enhancing local extremes

Extreme heat is one of the deadliest health hazards that is projected to increase in intensity and persistence in the near future. Here, we tackle the problem of spatially heterogeneous heat distribution within urban areas. We develop a novel multi-scale metric of identifying emerging heat clusters at various percentile-based thermal thresholds and refer to them collectively as intra-Urban Heat Islets. Using remotely sensed Land Surface Temperatures, we first quantify the spatial organization of heat islets in cities at various degrees of sprawl and densification. We then condense the size, spacing, and intensity information about heterogeneous clusters into probability distributions that can be described using single scaling exponents (denoted by β, $${{\boldsymbol{\Lambda }}}_{{\boldsymbol{s}}{\boldsymbol{c}}{\boldsymbol{o}}{\boldsymbol{r}}{\boldsymbol{e}}}$$Λscore, and λ, respectively). This allows for a seamless comparison of the heat islet characteristics across cities at varying spatial scales and improves on the traditional Surface Urban Heat Island (SUHI) Intensity as a bulk metric. Analysis of Heat Islet Size distributions demonstrates the emergence of two classes where the dense cities follow a Pareto distribution, and the sprawling cities show an exponential tempering of Pareto tail. This indicates a significantly reduced probability of encountering large heat islets for sprawling cities. In contrast, analysis of Heat Islet Intensity distributions indicates that while a sprawling configuration is favorable for reducing the mean SUHI Intensity of a city, for the same mean, it also results in higher local thermal extremes. This poses a paradox for urban designers in adopting expansion or densification as a growth trajectory to mitigate the UHI.

Statistics of Visual Responses to Image Object Stimuli from Primate AIT Neurons to DNN Neurons

Under the goal-driven paradigm, Yamins et al. ( 2014 ; Yamins & DiCarlo, 2016 ) have shown that by optimizing only the final eight-way categorization performance of a four-layer hierarchical network, not only can its top output layer quantitatively predict IT neuron responses but its penultimate layer can also automatically predict V4 neuron responses. Currently, deep neural networks (DNNs) in the field of computer vision have reached image object categorization performance comparable to that of human beings on ImageNet, a data set that contains 1.3 million training images of 1000 categories. We explore whether the DNN neurons (units in DNNs) possess image object representational statistics similar to monkey IT neurons, particularly when the network becomes deeper and the number of image categories becomes larger, using VGG19, a typical and widely used deep network of 19 layers in the computer vision field. Following Lehky, Kiani, Esteky, and Tanaka ( 2011 , 2014 ), where the response statistics of 674 IT neurons to 806 image stimuli are analyzed using three measures (kurtosis, Pareto tail index, and intrinsic dimensionality), we investigate the three issues in this letter using the same three measures: (1) the similarities and differences of the neural response statistics between VGG19 and primate IT cortex, (2) the variation trends of the response statistics of VGG19 neurons at different layers from low to high, and (3) the variation trends of the response statistics of VGG19 neurons when the numbers of stimuli and neurons increase. We find that the response statistics on both single-neuron selectivity and population sparseness of VGG19 neurons are fundamentally different from those of IT neurons in most cases; by increasing the number of neurons in different layers and the number of stimuli, the response statistics of neurons at different layers from low to high do not substantially change; and the estimated intrinsic dimensionality values at the low convolutional layers of VGG19 are considerably larger than the value of approximately 100 reported for IT neurons in Lehky et al. ( 2014 ), whereas those at the high fully connected layers are close to or lower than 100. To the best of our knowledge, this work is the first attempt to analyze the response statistics of DNN neurons with respect to primate IT neurons in image object representation.

Aggregate Demand and the Top 1 Percent

There has been a large rise in US top income inequality since the 1980s. We merge a widely studied model of the Pareto tail of labor incomes with a canonical model of consumption and savings to study the consequences of this increase for aggregate demand. Our model suggests that the rise of the top 1 percent may have led to a large increase in desired savings and can explain a 0.45pp to 0.85pp decline in long-run real interest rates. This effect arises from both a wealth effect at the top and increased precautionary savings from declines lower in the income distribution.

ESTIMASI NILAI VaR PORTOFOLIO MENGGUNAKAN FUNGSI ARCHIMEDEAN COPULA

Value at Risk explains the magnitude of the worst losses occurred in financial products investments with a certain level of confidence and time interval. The purpose of this study is to estimate the VaR of portfolio using Archimedean Copula family. The methods for calculating the VaR are as follows: (1) calculating the stock return; (2) calculating descriptive statistics of return; (3) checking for the nature of autocorrelation and heteroscedasticity effects on stock return data; (4) checking for the presence of extreme value by using Pareto tail; (5) estimating the parameters of Achimedean Copula family; (6) conducting simulations of Archimedean Copula; (7) estimating the value of the stock portfolio VaR. This study uses the closing price of TLKM and GGRM. At 90% the VaR obtained using Clayton, Gumbel, Frank copulas are 0.9562%, 1.0189%, 0.9827% respectively. At 95% the VaR obtained using Clayton, Gumbel, Frank copulas are 1.2930%, 1.2522%, 1.3152% respectively. At 99% the VaR obtained using Clayton, Gumbel, Frank copulas are 2.0327%, 1.9164%, is 1.8678% respectively. In conclusion estimation of VaR using Clayton copula yields the highest VaR.