Parameter: Journal of Statistics

CALCULATION OF ANNUAL PREMIUM RESERVES FOR INDIVIDUAL HEALTH INSURANCE USING THE PROSPECTIVE METHOD

Rido Evendi Tarigan — 2026-06-29

Premium reserve management is an important aspect of health insurance because inaccurate reserve calculations may affect an insurer's ability to meet future claim obligations. This study aims to calculate and compare annual premium reserves for individual health insurance under two types of insurance products, namely term insurance and whole life insurance, using the Prospective Method. The analysis was conducted through a simulation involving a family of three consisting of a father aged 40 years, a mother aged 36 years, and a son aged 9 years. The calculations were based on the CSO 1941 mortality table, an annual interest rate of 2.5%, and health insurance benefits covering Rp 200.000 per day for hospital room charges, Rp 75.000 per day for physician visits with a maximum of 180 days, and Rp 4.000.000 per treatment period for medical expenses. The results indicate that the annually renewable premium increases with the insured's age, with the total premium rising from Rp 704.060 in the first year to Rp 1.196.260 in the tenth year. The reserve values obtained using the Prospective Method show different patterns for the two insurance products. Premium reserves for term insurance gradually decrease over the policy period and reach zero at the end of the contract, whereas premium reserves for whole life insurance increase continuously from year to year. These findings demonstrate the different reserve characteristics of term and whole life insurance and provide useful actuarial information for insurers in planning and managing long-term reserve funds.

EXAMINING LIFE EXPECTANCY PATTERNS BY REGION, COUNTRY, AND GENDER IN AFRICA

Alfred Ayo Ayenigba — 2026-06-29

This study examines life expectancy patterns across 54 African countries, utilizing secondary data from the World Health Organization (WHO) and Statista. Employing descriptive statistics, a paired sample t-test, and a Wilcoxon signed-rank test, the analysis reveals pronounced regional disparities: North Africa records the highest average life expectancy (73.2 years), followed by Southern Africa (65.9 years), East Africa (64.1 years), West Africa (63.3 years), and Central Africa (61.9 years). A statistically significant female longevity advantage of 4.40 years (t = 19.18, p < 0.001; Wilcoxon Z = 6.1, p < 0.001) is observed consistently across all regions, with the widest gender gap in Southern Africa (6.2 years) and the narrowest in East Africa. Eritrea is a notable anomaly, where male life expectancy marginally exceeds female by 0.3 years. These findings underscore the need for gender-sensitive health policies and targeted regional interventions to address persistent health inequities across Africa.

ESTIMATION OF THE PERCENTAGE OF CRIME VICTIMS IN PAPUA REGION IN 2024 USING HIERARCHICAL BAYES SMALL AREA ESTIMATION

Muh Taslim Hakim — 2026-06-29

Reliable statistics on crime victims are essential for monitoring public safety and supporting evidence-based policies. In line with Sustainable Development Goal (SDG) 16, which aims to promote peaceful and inclusive societies by reducing all forms of violence, district- and municipality-level estimates of crime victims are particularly important for regional development planning in the Papua region. However, direct survey estimates at this level often suffer from low precision due to limited sample sizes. Therefore, this study aimed to estimate the percentage of crime victims at the district/municipality level in Papua Island in 2024 using the Hierarchical Bayes Small Area Estimation (HB-SAE) approach based on the area-level Fay–Herriot model. The model incorporated three auxiliary variables derived from the 2024 Village Potential Statistics (PODES) published by Statistics Indonesia (BPS): the proportion of villages with base transceiver stations (BTS), the proportion of villages with markets, and the proportion of villages reporting theft incidents. The results showed that the HB-SAE approach produced significantly more precise estimates than direct estimation. The mean Relative Standard Error (RSE) decreased from 47.35% under direct estimation to 37.02% under the HB-SAE approach, representing a 21.82% improvement in estimation precision, and the Page test confirmed that the reduction in RSE was statistically significant (p = 0.01). Furthermore, the HB-SAE approach successfully generated estimates for all districts/municipalities, including four non-sampled areas. These findings indicate that the HB-SAE approach provides reliable small-area estimates to support crime-related policy formulation and regional development planning in Papua Island.

EVALUATING THE ACCURACY OF HOLT'S DOUBLE EXPONENTIAL SMOOTHING AND ARIMA IN FORECASTING BATANG REGENCY'S GROSS REGIONAL DOMESTIC PRODUCT (2025–2029)

Alviyatur Rahmaniyah — 2026-06-30

Batang Regency has experienced continuous economic growth in recent years, making information on changes in Gross Regional Domestic Product (GRDP) essential for supporting regional development planning. Reliable forecasting methods are needed to provide an overview of future economic conditions based on observed trends. Previous studies have applied either Holt’s Double Exponential Smoothing (DES) or Autoregressive Integrated Moving Average (ARIMA) models for economic forecasting; however, comparative evidence regarding the performance of these methods on Batang Regency’s GRDP data remains limited. Therefore, this study aims to compare the forecasting performance of both methods and generate GRDP forecasts for the period 2025–2029. This study utilized annual GRDP data at constant prices from 2010 to 2024, which showed an increasing trend and nonstationary characteristics. The analysis procedure included descriptive analysis, stationarity testing using the Augmented Dickey–Fuller (ADF) test, model estimation, and forecast accuracy evaluation using Mean Absolute Percentage Error (MAPE). The ADF test indicated that the series became stationary after first-order differencing, and several ARIMA models were evaluated using the Akaike Information Criterion (AIC), resulting in ARIMA(1,1,1) being selected as the best-performing model. However, in terms of forecasting accuracy, Holt’s DES outperformed ARIMA by producing a lower MAPE value of 2.15%, compared to 8.87% for ARIMA(1,1,1). Forecasts generated by both methods indicate that Batang Regency’s GRDP is expected to continue increasing during the period 2025–2029. These findings demonstrate that Holt’s DES provides more accurate forecasts for the GRDP data used in this study and contribute empirical evidence regarding the comparative performance of classical time series forecasting methods for regional economic planning. The results may serve as a reference for policymakers in formulating more targeted and sustainable development strategies.

ANALYSIS OF THE RELATIONSHIP BETWEEN REGION, SEX, AND MARITAL STATUS USING A FOUR-DIMENSIONAL LOG-LINIER IN BENGKULU PROVINCE IN 2022

Putri Damayanti — 2026-06-30

This study analyzes the relationship among four categorical variables age group, gender, marital status, and region of residence among individuals aged 10–24 years in Bengkulu Province in 2022. The analysis employed a four-dimensional log-linear model to examine complex associations between these variables. Data were obtained from the Central Bureau of Statistics (BPS) and organized into a four-way contingency table representing all possible combinations of the observed factors. Model parameters were estimated using the maximum likelihood method, and model fit was assessed using the Pearson Chi-Square test and the Akaike Information Criterion (AIC). The results indicated that models involving only two or three factors did not fit the data adequately. In contrast, the saturated model that included all four factors provided the best fit, with a p-value of 1. These findings suggest that there are strong and complex interactions among age, gender, marital status, and region. Therefore, the relationships among these demographic factors cannot be explained independently but require a comprehensive model that incorporates all interactions.

PATH ANALYSIS : DIRECT AND INDIRECT EFFECT HYPOTHESIS TESTING USING JACKKNIFE RESAMPLING ON THE DEVELOPMENT OF A CASHLESS SOCIETY

Dea Saraswati Pramaningrum — 2026-06-30

This study investigates the direct and indirect effects of marketing mix variables, namely Product, Price, and Promotion, on the Development of a Cashless Society, with the Digitization of Electronic Money serving as a mediating variable. Path analysis was employed using secondary data obtained from a survey of 100 mobile banking users in Jakarta during the COVID-19 pandemic. To reduce dependence on distributional assumptions, hypothesis testing of the path coefficients was performed using the jackknife resampling method. The results indicate that Price and Promotion have significant positive direct effects on the Digitization of Electronic Money, while the Digitization of Electronic Money has a significant direct effect on the Development of a Cashless Society. Furthermore, Price and Promotion exert significant indirect effects on the development of a cashless society through the mediating role of electronic money digitization, whereas Product has neither a significant direct nor indirect effect. The proposed model achieved a Q-square predictive relevance of 0.5893, indicating moderate-to-high predictive capability. These findings highlight the critical role of electronic money digitization in facilitating the transition toward a cashless society and demonstrate that jackknife resampling provides a robust approach for hypothesis testing in path analysis when normality assumptions are questionable