This study examined how successful was the implementation of compulsory pension in Israel. There are different criteria for measuring the success of pension reforms, such as: the system’s financial stability; actuarial balance; examination of the market’s industry structure through changes in firms’ market shares; and examination of the quality and costs of investment management. An important measure of the efficiency of pension reforms is the extent of pension coverage in the labor market, and particularly the predicted replacement rate for the current workforce – a criterion which was at the forefront of this study. This study is exceptional in making use of actual data of the whole pension fund industry. This data is confidential and was analyzed for the first time. It was obtained by kind permission of the Capital Market, Insurance, and Savings Authority, which supported this study. Alongside this data, we also used administrative data provided by the Social Security services. The combination of these unique databases made possible the research which accentuates the data fed into the prediction model, so that the common assumptions used in most pension simulators were replaced almost entirely with actual data. This data refers to various characteristics of the labor market and individual savers:

• wage profiles (using estimated wage equations and simulations) were created, which estimate monthly wages of workers for each year of employment. These profiles were built separately for representative types of men and women at each monthly wage decile.
• Analysis of Social Security data presented for the first time the rate of insured wage out of total wage, differentiated by wage quintiles and gender.
• Analysis of pension fund data generated data on the rate of management fees paid per income level, rate of plan holders who draw their pension compensation , seniority of savers in pension funds, rate of funds allocated to pension and the insurance costs deducted, and more.
• Actuarial calculations based on an ideal model of continuous employment, pension coverage for the entire income, no compensation withdrawal and average management fees supposedly showed that the accrued pension allowance can reasonably substitute income during retirement years: the imputed replacement rate generated is 65% for men and 46% for women (OECD, 2016). Research findings which are derived from actual data, show that this prediction is wildly optimistic, since calculations based on actual data of occupational pension resulted in replacement rates of 35% for men and 29% for women. This data is population averaged. In the lower quintiles of the workforce, this problem is much more severe, and the inequality of pension coverage in this group is a major issue which warrants additional research.