This article follows up on themes discussed here and here by adding data about the association between location and price and income. The data source for apartment prices is a multi-year data set of the Israeli Ministry of Housing and Construction. The data set tracks the average apartment price over time in 74 localities with a total population of almost 6 million. This sample of localities includes all of the large cities plus an assortment of smaller towns. Arab majority towns are not represented.

Between the years 2008 and 2013 housing prices in Israel had increased significantly. The average apartment price rose from about 100 monthly salaries in 2008 to about 130 in 2013.

price-2008-2013Figure 1

The location-associated data shows (Figure 1) that the price increase can be described well as a proportional increase of about 30% plus an offset of about ILS 200,000. Read the rest of this entry »

A new version of this.

This has a revised Analysis section discussing the zero interest rate lower bound scenario and a new section showing a parametric example of the model including an illustration of the resulting functional relationship between the money supply and production.

The figure below shows the frequency of certain types of words in inaugural speeches of U.S. presidents. The three types showed are words beginning with the string “elect”, words beginning with “repub” and words beginning with “democ”. Around the middle of the 20th century the “elect” and “repub” types were largely abandoned while the “democ” type became more prominent.


The Data was extracted from the texts of the speeches as found on the website of the “UCSB presidency project” using an R script.

Algorithms for kids

February 15, 2014

I recently gave a talk about algorithms to a 6-th grade class. My slides are below (an English version and a Hebrew version).

What I believe worked very well was going through the execution of the algorithms step-by-step and having the kids take turns at saying in advance, before I switch to the next slide, what the changes they are expecting see are – where the instruction pointer arrow will move, what variables will change their values and what new output will be emitted.

algorithms-engEnglish version

algorithms-hebHebrew version

About learning

February 3, 2014

I have spent my childhood and adolescence implicitly believing in ideas about learning that in later life I have come to disbelieve. It appears to me that those ideas are quite common in our society, including among schoolchildren. These false ideas cause significant harm to many of those who believe in them and, being widespread, to society. I therefore think it is important to disabuse students, and particularly children, of these ideas.

The false ideas about learning maintain their hold in the mind of the public mainly because they are rarely examined. One of the false ideas is that the nature of learning is self-evident and needs no examination. This, of course, makes those ideas self-reinforcing.

This essay is a brief description of those erroneous ideas about learning and of an alternative view – in my opinion, the correct view. To make things concrete, I start by listing a set of practical implications of the rejection of the false notions and adoption of the alternative view. These are things that can and should be done by the various people involved in education – students, teachers, and managers of education systems.

This post presents some additional time series regarding the matters discussed in a previous post. In addition to using the data used in the previous post, one additional series is used – average rent. This is Israeli central bureau of statistics (CBS) series number 030000.




Apartment prices in Israel have gone up significantly in recent years. In the years 1996 to 2008, an average apartment cost about 100 average salaries. Within two years, this ratio has gone up to about 130. This rise can be explained to a significant extent by the reduction in interest rates, which implies that a given mortgage payment can repay higher loans. The interest rate on an indexed 20 year mortgage has fallen from about 6% in the late 1990’s and early 2000’s to about 3.5% in 2010 to 2012. As a result, despite the 30% increase in apartment cost to salary ratio, the ratio of the monthly payment on a 20 year loan for the cost of an average apartment to the average salary has gone up much more moderately – from about 70% to about 80%.


The fact that the increase in apartment prices is associated with a decrease in interest rates raises the possibility that a future rise in the interest rates (which would bring them close to historical rates) would cause a return of the apartment cost to salary ratio to its previous level.

Data sources:

  • Average monthly salary (nominal): Social security data (social security law, section 1), similar to statistical abstract 2012, Table 12.34, but more up to date, using the January 1st 2013 entry as reflecting 2012 data, etc.
  • Consumer price index (CPI): Israeli central bureau of statistics (CBS), series 120010
  • Average price of apartment owned by its residents: CBS series 020000
  • 20-year mortgage CPI-indexed interest rate: Bank of Israel data, using the “20-25 years” column for recent years and the “above 20 years” column for pre-July 2011 entries

Co-authored with Daniel Gat.


Note: The data here do not include settlers and construction in the eastern neighborhoods of Jerusalem, i.e., those areas of the West Bank that are included by Israel in the municipal boundaries of Jerusalem.

Data sources: Statistical abstract of Israel, various years; B’tselem report, By Hook or By Crook. Table 1 (based on Statistical abstract of Israel, various years). Data: 1, 2.

The increase of the Haredi (Jewish ultra-orthodox) population in Israel is a topic widely discussed in Israeli mass media and with considerable sway over Israeli politics both in the form of increasing Haredi electoral power and in the form of providing campaign and policy agendas for opposing forces.

Official statistics (social security annual report 2011, charts on pages 139, 140) indicate that the average Haredi family has about 6 children (with a decreasing trend over the last two decades) while non-Haredi Jewish families have about 2.5 children. Talk about demographic trends caused by the differential fertility rate, causing an imminent “Haredi  and Arabic” majority, is quite common. This is coupled with complaints about the fact that Haredi labor force participation rate is low, the fact that Haredim do not serve in the Israeli military and the fact that some Haredim receive stipends from the state as rabbinical students.

The chart below shows the proportion of votes received by Haredi parties in Israeli elections.


Data: text file, Google spreadsheet.