HtAG Platform Overview Video Pt.3

This is the 3rd and final video in the HtAG Platform education series. In this video we cover off:

  • Deep dive into the forecasts graphs
  • Learning how to read the GRC graphs for council and suburbs
  • Learning how to interpret LGA heatmaps & scatter plots
  • Interpreting the Demand Profile Tab
  • Conclusion to the Camden Council case study with 3 suburbs earmarked for investment

If you missed the previous two videos in this series, make sure to watch them before watching this video.

1 thought on “HtAG Platform Overview Video Pt.3”

  1. Video Transcript
    To recap, using the drop-down menu, we can pick and choose between different suburbs and see how areas have moved in the past and thus gauge the certainty of the forecast by recognising patterns in price movements. Looking at different graphs we can also aim to ascertain the relationship or the lag between number of sales and median price movements, which can assist in determining the impact current and forecasted sales will have on future price movements.

    In some instances, it is clear that changes in the number of sales does not always correspond to median price changes. One might ask why? Well because seeing increases in median price is dependent on more things than increases in sales. Although this realisation might seem self-evident, its benefits in relation to understanding market fundamentals for a particular area are rather not. To illustrate, when demand outstrips supply, it causes the median price to increase. However, if the graph depicts a downward trajectory for the price while its number of sales are projected to increase, it would be safe to suggest that there is an oversupply in the area and that demand is not strong enough to absorb the excess supply even though the graphs show an increase in the number of sales.

    Let’s take Spring Farm for example. It is forecasting an increase in median price. Its’ number of sales is decreasing respective to the increase in values if we compare the forecasted section of the graph with past correlation between number of sales and median value increases. Looking at the relationship between the number of sales and price increases in the forecasted section of the graph, it suggests that there is an undersupply of properties in that area.

    In this instance, if our investment window is 2 years, it can be said that Spring Farm is well positioned for individuals interested in “flipping” strategy. These types of investors can utilise the area’s dynamics to their advantage as the median value of an area is positioned to increase circa 30k on top of what the investor would make by adding value to the property.

    Median rent, on the other hand is positioned to decrease while number of rentals is set to remain steady and at high level. This suggests, that there is an oversupply of rentals or that there is not enough demand for rentals in the area, meaning the area could potentially be dominated by owner occupiers. This information, although implicit, can be easily obtained if one simply considers the past and forecasted relationship between sales and median price. This information is only derivable if we consider future performance in conjunction with past performance. This is another aspect that makes the forecasting feature so valuable, as it puts the past performance into context.

    Although at HtAG we aim to eliminate the need to obtain information on a plethora of other variables from other sources, we will just, for arguments sake, aim to vet the previous assumptions and look at ABS to see what the proportion of owner occupiers is to that of renters. Looking at the 2016 Census, Spring Farm has 22% of renters which is 9% less than average in NSW and 8% than the average in Australia. Returning the Spring Farm, it has an entry point of 680K which makes it the third cheapest area in Camden Council if we filter by median price.

    Its’ forecasted 2 year capital growth is 6.54%. Spring Farm’s rents are forecasted to decrease by about $10 dollars and Spring Farm is marked with high confidence. Using Camden Council as a case study, it can be said that Spring Farm could be the most suitable area to invest in based on the retrieved statistics in comparison to other suburbs in Camden Council. As rents are forecasted to decrease by only $10, investors will not have to dip in their pocket too much extra while investing in a suburb with 3rd highest forecasted capital growth with 3rd lowest entry point.

    One thing that we at HtAG cannot emphasise enough is telling our customers to consider all in formation in unison as looking at the graph alone will not tell you that the investment quality of Spring Farm is considerably higher than that of other suburbs in the Council, because of its low entry point and high forecasted capital growth. More importantly, looking at the relationship between the number of sales and rents and forecasted values one can also get an insight into market fundamentals and the relationship between supply and demand which would otherwise require us venturing to external sources and doing a lot of number crunching unnecessarily.

    Going to external sources means using ABS and other websites to collate information on several different variables such as unemployment, population, building approvals, migration, consumer sentiment etc to realise that Spring Farm provides opportunities for development. This is essentially a good example of our mission and vision—enabling every individual to become a property investor by simply using one tool while not having to be concerned by all of the complexities of investing using the traditional market research.

    The notion behind using a single variable such as median value and median rent to ascertain market dynamics has not come about involuntarily. At the onset of developing our service we conducted a regression analysis of up to 10 external variables and found that using a multitude of variables to model the property market can more often than not, have an adverse effect on our ability to forecast with high degree of accuracy.

    One significant discovery was that variables such as population growth and unemployment often have no correlation with price growth. On the other hand, building approvals – a variable that does have some significant influence – cannot be replicated as a benchmark for all other areas as each property market is unique. The second con against the multivariable model is that analysing such data to find meaningful correlation is such a daunting task that it can only be performed by a team of full-time statisticians with economical background. Furthermore, by the time they would have finished modelling the markets for all suburbs in Australia, half of their forecasts would be outdated due to influx of new data.

    It is due to the insights obtained from a regression analysis which demonstrated no pattern in correlation of different variables and median price, that we realised that a single variable, compounded with large amount of data points and market cyclicity regressors, is a much better springboard for projection of price movements and overall market dynamics.

    The best accuracy was achieved when we completely discarded other variables and only considered past sales data for a particular area and past sales data for areas in close proximity. As we conducted various tests and compared them to the actual performance of the suburb, we realised that this approach produces a model with the highest accuracy. So, how was than the accuracy of the model tested so that it provides a fixed confidence metric?

    The way our models are tested is that we produce a forecast for a particular moment in time in the past to see the difference between our forecast and the actual values. We travel back in time, so to speak, and pretend we are producing a forecast for the future. We than compare the forecasted values with the actual values that we already know have been realised and measure an error rate between the two values.

    We continually perform these tests to ensure the accuracy of the model. This guarantees that the error rate is always within the values that are provided for each respective confidence level listed on our website.

    So essentially, the model continuously improves itself and learns from its mistakes adjusting accordingly. One important thing to point out is that the more sales there are, the lower the error rate of the forecast will be. One can see this by simply looking at the number of sales for areas marked with high confidence.

    All areas marked as high confidence have a lot of sales recorded. This means that in the future, when there are data points collected for two, three or more cycles, the accuracy of the prediction model will increase for the 2-year projections. This will also enable us to project further into the future meaning that in 5 years from now, we may be able to project 3 years into the future with the same accuracy.

    The platform provides opportunity to benchmark different areas thus enabling customers to not only make a decision which is in line with one’s circumstance but also permitting one to vet suggestions and advices of other property professionals. Coming back to the graphs and forecasts, now that we have introduced a time dimension into the statistics, the provided data becomes more fluid and relevant.

    Having an adequate forecast in place permits our customers to better contextualise past performance enabling them to ascertain a disbalance in supply and demand having to examine only a handful of data points.

    This is an important distinction in respect to the report format presentational model which feels static as all it does is highlight numbers on a page at a point in time. The report model presentation becomes meaningful only when we compare reports of two or more different areas. The graph forecast, however, introduces a time dimension into our analysis because past performance now becomes more meaningful as it is superimposed against anticipated performance of a particular area.

    This is particularly evident for areas that have behaved in abnormal ways showing exorbitant increases in their median values followed by equally exorbitant drops. Such areas, although potentially forecasting substantial growth and capital gain, would be of low investment quality, and more than likely be marked with low or very low confidence, as their sporadic and unpredictable market movements would suggest week market fundamentals or an imbalanced supply and demand.

    Next, we move onto the growth rate cycle tab. It is important to point out that the growth rate cycle feature is a little different from what has generally been referred to as the property clock.

    The property clock notion suggest that each area follows a predictable pattern in its movements. When its median price has decreased to a point where it will not decrease further, it can be said such an area is at the 6 o’clock mark and has reached a bottom. Each bottom is subsequently and inevitable followed by a rise in prices to a point where the area’s median value does not rise any further suggesting it has reached its peak, that is, it is at the 12 o’clock position of the cycle. Understanding where an area is in its property clock cycle is used to time purchases and take advantage of the cyclical movements of areas.

    The strategy of timing purchases becomes more pertinent if one’s investment window is one cycle or less. If my investment window is two years, for example, then the property clock really needs to be taken into consideration because if I’m buying into a declining market, prices will decline even further minimising my ability to leverage any growth into additional purchases.

    So, if my window is two years and I’m buying in a declining market I will experience further reductions in my return on investment. This concept of timing your purchases is the main philosophy behind the property clock notion that has been used by property investment professionals for decades. The growth rate cycle provided by HtAG is a little different. While the property clock refers to the cyclical movement of property values, the growth rate cycle considers the percentage of growth change as opposed to the change in value of a property. Looking at Camden Council, it has grown 5% in 2017, while it grew 10% in 2019. This means that the growth rate cycle feature tracks the changes in the percentage of growth year on year.

    And as you can see, same as the property clock, the growth rate movements are also cyclical.

    Looking at Camden, we can see that the Council’s growth rate moves in a cyclical fashion as all of the suburbs in the council, each represented by the individual line, follow a similar trend that has a bottom and a peak and then a bottom again.

    We can see that in 2009 the area reached the bottom in its rate of growth which is the same as saying that the area was growing at a zero % rate or below. 6 years later, in 2015, the area reached its peak growing at approximately 12 to 13%. Subsequently, its growth rate began to decline to now be in the negative territory which was also highlighted in the ranking table at the onset.

    The red line on the graph signifies the zero point. Everything under the red line is negative growth. Everything above red line is positive growth. It is important to point out that the growth rate cycle feature is much more sensitive and nuanced than the property clock. We can see by looking at the graph that areas also move by following a macro cycle that spans from 2009 to today.

    However, they also have micro movements or micro cycles. We can see by observing a particular line representing a suburb that in 2012 its growth rate was 7.5%. In the next couple of years its growth rate increased to 15% subsequently falling to 13.6%. As we can see, the areas have experienced a mini cycle within a period of a couple of years in addition to the macro cycle of spaning 10 years and over.

    Having a more sensible understanding of the cyclical movements of an area becomes very pertinent to customers with shorter investment windows. Understanding that areas actually follow two different patterns in movement simultaneously depending on the scale of interest, although counter intuitive, is very important for anyone who times the market. As we can see, on a macro scale area follow a more or less predictable and steady patterns in movement however when we look closer, on a micro scale all move in different way respective to one another.

    This is best seen when one does a comparison between different areas. It becomes evident that while one experiences a peak, the other is at the bottom, which becomes extremely important when one benchmarks suburbs for their investment quality. As mentioned before, all else being equal, counter cyclical investing, that is, timing the market to buy at the bottom, is a very important property investment strategy.

    This cannot be determined from a property clock, or at least it cannot be done in such a nuanced and sensitive way. For example, the property clock suggests that when an area is at the bottom of its cycle, it is positioned for its median value to increase subsequently. For example, if the area hits the bottom at 200k in its median value, being at the bottom would mean that its value is positioned to increase.

    However, what the growth rate cycle suggests is that that’s actually not the case at all. An area that is at the bottom of the cycle or even an area that is in the rising part of the cycle can still experience a decrease in its median value. To make it simple, areas can turn a corner but still be experiencing negative growth, although at a reduced rate than it did at the bottom. This also is true in the inverse scenario—an area can have a declining growth rate but be experiencing increases in its median value.

    So an area which is at the bottom, or an area that is rising can still experience decreases in median value.

    Take Camden for example. Looking at the graph we can see that Camden was experiencing a decline in its growth rate between 2016 and 2019. However, its median value was still increasing, in 2017 for example, at a rate of about 6%. This means that although Camden is in the decline stage of the cycle, its median price has growth by 6% in 2017.

    The same happens when the area, in this instance Camden, has reached the bottom. We can see that the area is experiencing negative growth since the line is below the red line which is indicative of zero. Being below the zero line means that its median value is experiencing negative growth which means it is decreasing. However, its negative growth is slowing down or reducing which means that although the median value is reducing further, the area has turned a corner and is very soon positioned for an increase in its median value or positive growth. So, the growth rate cycle actually provides you with information of what’s going to happen before the area rebounds.

    So, before you actually see increases in median value, the area has to rebound from the negative growth. If we were considering timing the market in Camden for example, we would purchase in 2021 because the negative rate of growth for Camden will slow down at that point positioning the area for a rebound.

    This is why the property clock actually becomes obsolete if one takes into consideration the GRC. Looking at this page again, one is provided with options to filter through different suburbs in Camden or to choose the GRC for Camden Council as an overall. It is essentially a visual representation of the statistical information we saw at the home page when we searched for Camden.

    When we looked at the ranking table, we suggested that Spring Farm looked like the most favourable investment. If we look at the GRC, will we change our opinion? Looking ta the graph, Spring Farm was at the bottom in 2019-2020.

    The area has clearly rebounded; it is past the negative growth line, which means it has already been experiencing positive growth. This mean that everything from this point on in near future will be positive growth. Looking at the overall cycle of the area including all of the previous statistics, it can be said that Spring Farm ticks all the boxes if we are interested in investing in Camden. We do not see it going into negative growth due to the behaviour of the curve in the past and we anticipate it peaking again in the next 5 years or so.

    If I was to penny pinch and try and invest in an area that is at pure bottom ensuring than all positive growth experienced is pocketed, Narellan Vale could potentially be an even better investment. This is because Spring Farm has already experienced some growth which we have not harvested if we enter now.

    While Narellan Vale is yet to come to a point where we can harvest positive growth which is looking at the forecasted section of the GRC graph is positioned to happen in 2022. Going back to the ranking table, it is evident that Narellan Vale had one of the lowest entry points which means that it will be relatively easy to enter this market while also experiencing a reduction in its negative growth for rents. More importantly, it is also marked with high confidence.

    This is a perfect example of our mantra that to make a good investment decision, all features should be considered in unison. Spring Farm is a good investment, however if we consider the statistical information together with the growth rate cycle, Narellan Vale comes out as a new contender.

    Considering these features in terms of the professional subscription, one is provided to use them to benchmark all suburbs and Councils in Australia. You can search away based on the amount you are prepared to invest and other personal criterium to shortlist a number of suburbs that can be benchmarked using all features mentioned in unison.

    Now that we’ve explained this, the next feature that we would move on to is heatmaps.

    Looking at heat maps alone can give you a lot of information about existing hotspots as well as any potential spill-over areas. By spill-over we mean areas that are side by side to an area that has peaked and as a result is becoming too pricy. The theory is that this subsequently increases the appeal of the neighbouring areas. This leads us to another way we could analyse heatmaps—namely by identifying growth cluster and growth corridors with a particular Council or more Councils combined.

    When we move onto the heatmap page, we can see all suburbs in Camden represented on a map with each suburb having a variation of colour red and green. The red signifies negative year on year growth in the current quarter, not forecasted growth. We will take this opportunity to mention that we are working on a feature where we will have a heat map which also includes forecasted growth, that is going to make it easier for our customers to evince spill-over suburbs. In the meantime we suggest to download one of the Tableau reports from the store to get access to multi-year heatmaps.

    Referring back to the map, there is only one area that is green meaning only one area in Camden Council experiencing relatively higher year on year growth when compared to other suburbs in this council.

    The area is Catherine Field. By clicking on the information pin, we can see that the area has not had that many sales indicating its high year on year growth should be taken with a grain of salt. This area would be ranked with low confidence meaning that making any investment decisions in terms of its year on year growth and projected growth is risky.

    If we looked at the capital growth alone, this area would be more than suitable. However, if the previously covered metrics are taken into account our decision will be swayed elsewhere. For argument sake, lets look back at other pages and features to see how Catherine Field is behaving.

    Looking at Catherine Field in the ranking table, our intuition comes true and as we can see the area is marked with low confidence.

    The projected capital growth is going to be five percent. This means that the area’s exorbitant year on year growth is forecasted to grow even further. Rents are also forecasting an increase. Referring to the area’s entry point which is circa 1.7 million, those who can afford it would be enticed to invest in Catherine Field because of its forecast growth since 6% of 1.7 million is nearly over 100k return in 1-2 year time frame. Big return.

    But than if we look at the forecast page, which is a graphical representation of the data tabulated in the ranking tables, we can see that the market fundamentals are not in balance since the area shows sporadic movements in its median value and rent changes.

    We really do not need to refer to ABS or other data providers to find out that either the area’s population is not growing at the right pace, or its unemployment levels are out of balance or that there is an overall low demand for the property in the area. All I need to do it look at the graph and see that the behaviour of the curve suggests something is out of balance.

    In case of Catherine Field, following the growth statistics alone would be a very costly wrong decision. We would consider this area as an outlier in Camden Council and discard it at this stage.

    Returning back to how one can interpret the heat map, we encourage customers to look for cluster and corridors where there are a few areas connected with good potential. This means that irrespective of where within a cluster you buy, the area is positioned to benefit from growth spiling in over from adjoining suburbs.

    This area in particular does not nuance any clusters, meaning that the entire area moves in a more or less unitary way. Although this Council does not seem appealing due to the fact that it is in negative growth territory, its’ heat map arrangement, tells us that when the Council turns a corner, all suburbs within a council will exhibit a similar growth pattern.

    Realising this one can go back to GRC to see how Camden as a Council is behaving overall.

    Looking at the GRC page, we will see that Camden Council Is positioned to rebound at about 2021. This means that post 2021, nearly all areas within the Council will experience positive growth at different rates which serves as a good planning tool.

    And why is this important? Well, since different suburbs within Camden Council have different price points, one can strategically plan and enter those suburbs which are closer to current hotspots, placing them in a good position to take advantage of increased demand in the area as a result of now overpriced adjoining suburbs.

    Drilling further into what the heat map page has to offer, although in case of Camden one cannot find a particular sub-cluster of suburbs, the heat map has a scatter plot option which provides customers with ability to find clusters of streets in high demand within a particular suburb or between different suburbs.

    The scatter plot highlights where the sales have been concentrated the most. As such, looking at the scatter plot, we can see areas that are more popular due to the higher concentration of the number of sales made. Overall, we can see three different clusters within Camden Council if we look at the scatter plot as opposed to the heat amp.

    When we try to ascertain which suburbs, these identified clusters belong to we can see it is Spring Farm and Narellan Vale. Not surprisingly we identified these 2 suburbs as areas with the highest investment quality when we were explaining previous features. The other two clusters would be Harrington Park and Oran Park, which could also be in contention.

    The good thing about this feature is that we can see which areas are more popular than others which adds another layer to previously provided statistics and graphs. Investors would need to focus on these three clusters and work out how the statistics from these areas fit into their personal circumstance.

    This is where we come to the last tab on the LGA page. After ascertaining areas and clusters of interest, are ready to shift onto the demand profile to determine which dwelling type is mostly sought after within this council area. On this page we can see which dwelling type dominates a particular area as well as the composition of the dwelling that dominates, i.e. number of bedrooms within the dwelling.

    The y axis on the bar chart shows the number of sales. As such, what we see is information on which dwelling type is in the highest demand, or better yet, which dwelling type is the most popular. So, if we had to make a decision based on the information provided for Camden Council, we would look to purchase a 4 bedroom house as our next investment. Going back even further, given the information we obtained from the ranking table, the forecast, GRC and heat map tabs, we would invest into a 4 bedroom house in Spring Farm or Narellan Vale.

    Understanding what the most popular dwelling type is, will also assist in mitigating the risk associated with loss of rental income specifically if the areas chosen are forecasting a reduction in rents within the next 2 years. This means that buying the most sought-after dwelling type might eventuate in no loss or reduction in rent as the ranking table and the forecast page could have suggested. We do however suggest that customers filter based on the number of bedrooms also at the ranking table page to confirm this assertion for their area of intertest.

    This page can also be utilised when assessing the value of a single property one intends to purchase. Looking at the median value of a particular area and the demand profile we can approximate a reasonable value of a specific property; or at least reasonable to the point of matching desired investment statistics.

    We would like to take this opportunity to announce that HtAG is working on a service which will provide customers with the ability to forecast values of specific properties as opposed to localities only. For now, however, this can be approximated from the existing features provided.

    Using this case study as an education tool, where would we invest in given the insights we received from browsing through the statistics provided for Camden Council?

    The recommended approach would be to invest in one of the three clusters highlighted in the scatter plot. We would then proceed to rank these three localities based on their position in the growth rate cycle with a caveat of them having to be either at the bottom or juts passing the bottom of the cycle to maximise ROI.

    Given the GRC, if one wanted to time the purchase, year 2021 would be the right point of entry for all of the three areas in question. Investors should target a four-bedroom house for their purchase.

    This highlights a top down approach to analysis that not only serves as an educational tool but permits our customers to benchmark many areas in respect to their financial circumstance and investment strategy.

    This tool can be used doe better decision making at any point in one’s investment journey. Some of our home buyer customers purchase a subscription at a point when they are settling on a property to perform one last sanity check.

    More importantly, the tool becomes extremely useful to those who have different investment window timeframes. A two-year investment window will introduce different criterium to that of a ten-year investment window. Fir instance, if one is looking to renovate and sell, it would be paramount for him or her to find an area which has passed the red zero growth line and has rebounded in a way that it is experiencing positive growth. This ensures that the negative growth does not eat into the property’s added value.

    Moving on from the specific pages one can find other information on our website. Customers can have a read of our values, mission and vision. This information will tell you why this service was developed and what our future goals are. The website also provides information on different pricing structures and what services are included within different subscriptions. Our subscriptions are monthly based however we have provided customers a considerable discount should they opt in for yearly subscriptions.

    The frequently asked questions page provides answers to all of the questions asked over the years. This can be a useful starting point if customers are struggling with understanding how we differ from other providers and how we have managed to merge the old and the new and integrate data science into property investing. Moreover, the FAQ page does provide an exhaustive list of answers but if you are still unsure about something, please feel free to contact us at any time.

    Thank you for listening to this three-part video. Happy investing.

Leave a comment