Monopoly Money

Over a recent game of Uno, a friend commented on their competitive streak when playing Monopoly (triggered by me cheating against an 11 year old). Monopoly as a game doesn't really appeal to me as my competitive streak lasts all of 15 minutes, or until the bowl of game snacks runs out and it doesn't really help that there isn't a readily identifiable end point. However, it did remind me of some research I did out of curiousity a few weeks ago.

I had wondered whether or not the prices on the Monopoly board in anyway reflected the prices of properties in London in 1935 when the game was first published and whether the differences in value between properties was representative of real house prices at the time.

Clearly, without the benefit of particulars from real estate agents of the period it's difficult to know whether they were representative.

Step 1: What are properties worth nowadays

What I have done is investigate the average sale price of properties on Monopoly Streets in recent years. Unfortunately the British publishers of the game didn't give due regard to someone investigating these questions 77 years down the line. Some of the problems I have faced in undertaking this task have included the fact that a number of properties do not have residential properties on them (Trafalgar Square, Leicester Square etc), or are very short streets (Vine Street is all of 30 metres long) meaning there is low turnover of properties and thus available data. For others like The Angel, Islington, which aren't even streets, I've used Upper Street as a proxy (seeing as Pentonville Road is already included).

Of the 22 monopoly properties, I have managed to research indicative prices for 14 of them using Land Registry data. In addition I've also used estimates from Zoopla which predict values based on other local properties and historic sales figures. This is however somewhat of a data parlour game rather than rigorous research and analysis.

Monopoly inflation

So,using the original price per the cards, and the recent sale proceeds I calulated the effective rate of annual inflation for the properties is 10.2-12.0% based on the average value of sales and 11.1-14.0% based on Zoopla's estimates (assuming the Monopoly prices are realistic!).

Now, the question is, is whether or not this is a realistic level of house price inflation - if it's not, then the Monopoly prices are clearly not representative.

Measuring house price inflation is very difficult - there are all sorts of issues, such as the mix of property, the number of sales and the types of properties coming to market. I've looked for an index which goes back as far as possible.

Nationwide's index goes back to 1952, and estimates that the average house price in 1952 was £1,981, with the average for older houses calculated at £1,524 and 'new houses' as £2,107. From 1952 to 2013, the effective annual inflation rate was 7.58% for all properties and 7.98% for older properties, which is 2-6 percentage points less than our Monopoly inflation index, which means unless there was a large amount of inflation from 1935-1953, it looks like the prices in Monopoly are too low.

However, the Nationwide figures, are, national, and don't consider local variation. Halifax's index doesn't go as far back (only 1983) but does calculate regional inflation, with London inflation 17% higher than the UK as a whole. Applying this uplift to the Nationwide figures still however only lifts the inflation rate to 9.3%.

Playing around a bit more, if we use our inflation rates based on modern prices against Monopoly card prices to inflate the prices until 1952, and then uplift using Nationwide's index from 1952 to 2013, the modelled values are markedly lower than actuals. So all in all it seems that even in 1935 £200-300 for a property was actually a bargain.

But, what is interesting is the relative pricing of the squares in 1935 compared to the relative pricing in 2013. If you look at the chart below, you can see there is actually a reasonable correlation between the monopoly prices (x axis) and the actual prices (y axis). The two outliers are Bond Street (£320/£700k) and Fleet Street (£220/£392k), both of which I expect are affected by the small number of residential properties changing hands. So even if the prices weren't accurate in 1935, they appear to be relatively accurate for today's property market.

What price for a gold medal?

As I eat my lunch, I often have a browse of the FT data blog, ONS, data.gov.uk or the London datastore. Last week I had a look at the data on Olympic tickets sales and wondered what if any interesting patterns in the data there were.

I decided to build a little model to understand how different factors affected the average price of a ticket sold to the public. Unfortunately the data is incomplete in some sports (boxing, wrestling and gymnastics) as there hasn’t been full disclosure on the relative number of tickets sold to the public rather than corporate sponsors and for some events (road cycling), the comparison is affected by possible attendance for non-ticket events.

So the steps to estimate the cost of the average ticket coming out of the model are:

(A)  multiply the number of people attending (proxy for capacity) by -0.000044

(B)  multiply the % of tickets sold to the public by -1.18

(C) Decide if you want to see a medal event, if so, add £65.

(D) Choose your sport from the table below

(E)  Predicted ticket price is calculated by A+B+C+D

Athletics	£  220.26
Cycling Track	£  145.15
Diving	£  168.33
Equestrian	£  141.69
Hockey	£  129.91
Swimming	£  148.11
Synchro	£  123.92

This is obviously a very simplistic model that only considers 4 factors – capacity of stadium, how much is given to the public, whether a medal would be won and the sport itself. The results seem quite intuitive – but there is an issue in that sports are only played in one stadium, and the capacity may be partially determined by either the assumed popularity or demand for tickets, or the physical constraints of the space. Comparing an event in stadium with 10,000 seats compared to one with 80,000 seats only affects price by £3, so this doesn’t seem to be a significant factor in its own right. Ultimately the analysis is useful, but direction of causality is difficult to assess.

What is perhaps more interesting is when looking at some descriptive statistics - in this case the average price of a public ticket in different sports and then looking at the premium you would have had to pay for seeing a medal being won - basically you're looking at paying at least 50% more on average and the average across all sessions of sports was £93.

	No medals	Medal	Premium	Medal mark up
Athletics	87	235	148	170%
Cycling Road	13	24	12	96%
Cycling Track	71	164	93	132%
Diving	85	154	69	82%
Equestrian	56	104	48	85%
Hockey	41	70	29	72%
Swimming	67	167	100	150%
Synchro	47	72	25	53%

And now for a message for our corporate sponsors:

Looking at the number of extra tickets that sponsors and dignataties received, we can see that there was a greater number (and proportion) of tickets given to 'clients' at medal events. In the non-medal road cycling events, only 1 per cent of those with tickets (ie in a grandstand) were corporates, but this jumped to 25% in the medal events. In the main stadium, an extra 7,931 seats were taken up for sessions where medals were won by corporate guests.

I appreciate the need for corporate sponsors and dignataries to be given tickets, but when one can see that in track cycling and swimming more than half of the spectators were there because of who they were.
 

	Non medal event	Medal event	Extra tickets to corporates
Athletics	15,203	23,134	7,931
Cycling Road	59	766	707
Cycling Track	2,179	2,264	85
Diving	2,719	3,677	959
Equestrian	3,893	5,109	1,216
Hockey	2,880	4,919	2,038
Swimming	4,589	6,090	1,501
Synchro	2,476	3,059	583

 

	Non medal event	Medal event	Additional % of crowd on a jolly
Athletics	24%	37%	13%
Cycling Road	1%	25%	24%
Cycling Track	51%	54%	3%
Diving	27%	36%	10%
Equestrian	19%	27%	8%
Hockey	21%	35%	14%
Swimming	39%	55%	16%
Synchro	20%	24%	5%

One final thing - 76 sessions had no medal events and 46 did- this is a split of 62%:38%, but when the total revenue generated is calculated - 64% of all revenue from these events came from sessions with medal events, and half of all revenue from the above events came from athletics medal events, so charging a premium, clearly helped out financially. Others might ask if non medal events were subsidised, though the accessibility of the games to paying members of the public is another matter.

Population growth - to infinity and beyond?

 Projections of population growth have recently been released for London boroughs with projections extending to 2041. Clearly, these are just projections, but I assume those responsible know what they're doing.

If we focus on LBTH's population, you can see that between 2001 and 2041, it is expected to double from 202,000 to 400,000, and half of this increase will have occured by 2017.

The population estimate for Tower Hamlets in 2013 is 270,000, which shows a quite substantial increase in 12 years.

Using this data we can calculate a compound growth rate, which for Tower Hamlets is 0.72% per year. However, if we look at the year-on-year increases in population, which I've plotted below, we can see that the largest proportionate increases are expected to have already occured between 20004 and 2011, with an gradual stepped decrease in growth rates indicating that the growth rate is expected to slow.

 However, the actual number of new residents will still be reasonably significant and will increased by between 7,000 and 8,000 each year until 2021, before dropping to 5,000 new residents per year until 2026 and then 3,000 per year until 2041. All of these new residents will clearly drive a need for improvements in transport, health and education infrastucture.

But, is Tower Hamlets different to the rest of London? The simple answer is yes. I've indexed the forecast populations of LBTH, Inner London and Outer London in the chart below to show the relative level of population growth. Whilst LBTH's population is expected to double (100 to 200 on the chart below), Inner London's population will only increase by 46 per cent and outer London by 32 per cent, so whilst all of London will be getting even more crowded, we may feel more of a squeeze in Tower Hamlets than elsewhere.

To contexualise this, in 2001, the population density was such that if everyone stood outside, and spread themselves out equally into a grid, there would be one person every ten metres. In 2041, this will have reduced to one person every 7 metres (and as a result each square actually halves in area from 106 square metres per person to 54 square metres a person).

Reservoir levels - Lee Valley 1987-2012

Part of an on going attempt to improve the quality of my data visualisations on the blog, I'm attempting to come up with slightly more innovation or interesting ways of presenting data graphically. This chart shows the amount of water in the Lee Valley Reservoirs between 1987 and 2012.

Read more »

Wapping in numbers: Community safety

Creating maps in Excel pivot tables

When I can't be bothered doing something because it's too much hassle, I like to find a work around and data analysis is no different. When I was looking to plot fire brigage data onto a map, I faced having to convert 23,000 coordinates from the Ordnance Survey National Grid into the cartographic map projection used by google maps if I would be able to plot them without buying some fancy pants GIS software. Using my knowledge of MS Excel, I was able to achieve exactly what I wanted to in less than five minutes.

Read more »

Modelling fire brigade responses

In my last post I explained some of the processing of the London Fire Brigade incident data. In this post I look at the actual data and begin to consider what the impact of the proposed closure of Bow Fire Station.

Read more »

Mapping the impact of fire station closures: methodology

This post sets out my methodology for modelling the potential closure of fire stations. The actual analysis will be included in my next post.

Read more »

Language of Learning

To try and understand the 'Nappy Valley' phenomenon, I have been looking at data on the schools in the area. One of the data fields I have, is the proportion of school pupils for whom English is not their main language.

From the census I know that 34% of residents in Tower Hamlets do not think of English as being their main language, so I was a little surprised that 77.5% of pupils in Tower Hamlets state primary schools do not speak English as their main language- this is over twice the level of the general population. In St Katharine's and Wapping ward in 2011, 72% of the population speak English as their main language, yet in the schools (albeit a slightly wider area), only 20% of children speak English as their main language.

Read more »

Fire Station Closures

With the possibility of fire stations being closed, I thought it would be interesting to look at some of the issues that may be of concern to residents of Tower Hamlets.

The data I'm using is the London Fire Brigade's incident data for the period 2009-2012. For Tower Hamlets, this is just under 23,400 lines of data (one line per incident).
Read more »

Primary Schools for the Wapping catchment area

The first post in a series on the relative performance of schools in the local area. Here I begin by identifying the relevant primary schools I'll be looking into. In subsequent posts I'll look at their performance before moving onto secondary schools and borough-wide educational issues. Depending on availability I will include information on independent schools, but my focus will be on state funded schools.
Read more »

Nappy Valley 2

I was looking at the population structure of Wapping compared to LBTH, and London. I thought it would be interesting to identify other wards with a similar demographic structure to Wapping to see what shared characteristics there might be as part of my quest to understand the Nappy Valley situation.

For all London wards I calculated the proportion of the population that was a school aged child (5-19). A few statistics of note:

• Taking the average of 625 wards gives an average school age proportion of the population to be 17.1%
• The median proportion of the population that are school age is 17.4%
• If you take the population of London as a whole, 17.2 percent of the population are school age.
• The highest is 27.8% (Fieldway in Croydon)
• The lowest is 6.8% ('West End' in Westminster)
• Wapping at 9.5% is in the second percentile (i.e. 98 per cent of wards have proportionately more children)

Wapping is clearly at one extreme, so I filtered for all wards in London with fewer than 10% of the population of school age. This gave me 19 wards in 7 boroughs. The list and the proportion of the population of school age children are shown below:

City of London	All Wards	7.4%
Camden	West Hampstead	10.0%
Kensington and Chelsea	Earl's Court	9.2%
Kensington and Chelsea	Hans Town	9.8%
Kensington and Chelsea	Royal Hospital	9.9%
Merton	Hillside	9.4%
Southwark	Riverside	9.9%
Southwark	Surrey Docks	9.5%
Tower Hamlets	St Katharine's and Wapping	9.5%
Wandsworth	East Putney	9.8%
Wandsworth	Fairfield	8.1%
Wandsworth	Shaftesbury	8.5%
Westminster	Bayswater	9.0%
Westminster	Bryanston and Dorset Square	8.8%
Westminster	Lancaster Gate	8.3%
Westminster	Marylebone High Street	8.1%
Westminster	Tachbrook	8.7%
Westminster	Warwick	8.5%
Westminster	West End	6.8%

However, just because these areas all have relatively few school age children doesn't mean they're similar - it could be that they have a very large number of pensioners for example, skewing things, so I plotted a chart:

There appears to be a strong visual correlation between these wards with perhaps the exception of 'Hans Town' and 'Royal Hospital' which show less of a spike in young adults, but do show the same lack of children (as is expected given the basis for selection). However, looking just at the school age section, we can see these wards are all tightly packed, though some don't show the same steep drop.

I tested the goodness of fit of Wapping against this subset using categories of 5-9,10-14, 15-19 and 'other' to see if the profile of young people is consistent. This gives a p value of 0.063, which in simple terms means that at the 95% confidence level that these other wards are a good fit with Wapping.

I also looked at what I will refer to as the 'drop-off', that is, the percentage drop in the number of the population aged 5-9 compared to aged 0-4:

• Median drop-off is 17.9%, average is 17.2
• Wapping is in the top 1.3 percent in London with a drop off of 45.5%
• Looking just at Tower Hamlets, there are 3 Wards - Wapping and the two Isle of Dogs wards, which have a much greater level of drop-off

So there appears to not only be fewer children, but that there are fewer school age children compared to the number of toddlers.

	Drop-off
St Katharine's and Wapping	45.5%
Millwall	40.9%
Blackwall and Cubitt Town	33.9%
Bow East	23.0%
Spitalfields and Banglatown	18.3%
Bow West	17.7%
Limehouse	17.6%
Shadwell	17.5%
Whitechapel	17.0%
Mile End East	10.8%
Weavers	9.6%
Bromley-by-Bow	9.1%
Bethnal Green South	8.8%
St Dunstan's and Stepney Green	7.4%
Mile End and Globe Town	3.3%
Bethnal Green North	3.2%
East India and Lansbury	-2.7%

So next stage is to start understanding what it is that is a) driving the exodus and b) pulling in the youngsters, and then find out if the wards I have identified have anything in common - looks like I best get some datasets out!