Last time we looked at Indexation and Inflation and how it can impact on a business and its ability to remain profitable in multi-year contracts or indeed to have in place mechanisms where contracts become delayed to ensure the value remaining is indexed.
We saw how the main inflation measures, Consumer Price Index (CPI) and Retail Price Index (RPI) have over the last 2 years been extremely low, in fact historically low and that the previously depressed wage measure, Average Weekly Earnings (AWE) has moved in the opposite direction and wage inflation and its implications are now a significant factor for business in trying to manage their costs.
The Indexation Model that this post will relate to is subjective and cannot be used across the spectrum. However the basis of the Model is sound and it can be adapted to suit the environment in which it is going to be used. If we were to define what a Indexation Model should do, it would be as follows:
An Indexation Model takes past economic cycles to model expected future economic cycles to negate the Time Value of Money impact of those events.
In plain English it means you have a method of ensuring that when you raise the value of your contract for inflation it’s reflective of the pressures that your industry faces and is not a random measure like CPI and in the long run will ensure that your expected profit margins are able to be maintained. This is on the assumption that you manage all other cost pressures.
Despite what we have just said about CPI or RPI, the actual starting point of the Indexation Model will be one of these measures as what we are trying to do is take general inflation and then industry specific inflation data to determine the additional indexation, expressed as a percentage required to maintain our expected level of profitability.
Economists of the world talks about economic cycles and the general consensus is that over time the peaks and troughs even themselves out as effectively self-re-adjustment takes place. This of course is true in stable countires and this form of Indexation Model could not be used where despots have destroyed the economic system, such as Venezuela and Zimbabwe as here we have to add the currency effect.
We could argue that the current depressed CPI, but higher AWE is a re-adjustment that is playing out within the current economic cycle. That is why not utilising an Indexation Model that is empirical in nature can cause a business problems. The snap shot in time over over a short period shows low inflation and higher wage inflation. While this is acually a realistic portrail in the long run, its skewered currently as AWE has been supressed for the last 6 to 7 years.
The Indexation Model that I developed needed to apply an Indexation factor above RPIx. This is an enhanced measure of RPI which strips out the effects of Mortgage Interest. This gave a trend line over 34 years for RPIx alone (See Chart Below)
Then all of the factors that influenced the real indexation of the project were modelled as a percentage of the overall spend, this produced another trend line, based on their performance over time. (See Chart Below)
As can be seen, some of the commodities that are required to deliver the works have gone through periods of real peaks and troughs. How the impact of these peaks and troughs were modelled will be demonstrated later in this post.
The variance between the two gave the extra over percentage that was required to maintain parity, which gave the equation:
Indexation Factor = RPIx +x%.
So how far back do we go to model the future? This is the Marty McFly conundrum.
There is no right or wrong answer and different data sets are obviously going to give different results, as if we used data for the last 10 years only a significantly different result would be formed. However I would say that the minimum should be 3.5 economic cycles.
Which asks the question, “how long is an economic cycle?”
Ask 10 economists and they will probably give you 10 answers. So I have assumed an economic cycle is between 7 and 9 years (some are shorter, some are longer, but they average out at this length of time) and my base model used a data set of 34 years, which included the tail end of the mega inflation that the United Kingdom in particular (and western world in general) suffered in the 1970’s, as a result of weak government and volatility around the supply of oil.
Using the base year at which the rebase of the indexation began of 1976, where this had the notional value of 100 points, this resulted in the plotted trend line for RPIx that was shown earlier in this post.
In order to model the effect of all of the inflationary pressures that were expected to be encountered, a “basket” of goods” was established which assumed that for every £1 spend the make up of that would be according to the table below.
As the model was relatively reliant on labour (both skilled and non-skilled) for which the Average Earnings Index (AEI) was used. This has subsequently been replaced by AWE, so there is some skewering of the data at the cross over point. As the balance of the “basket of goods” are aggregates, cement based and oil based or derived products, the Baxter Indexes (Collated by Building Cost Information Services (BCIS) of the Royal Institute of Chartered Surveyors (RICS) were used for the following measures and with the following percentages attached (See “Cost Category Split” above) In effect the percentage and measure said that for every £1 of base turnover (no overhead and profit implication) each item was made up the following percentages (or Pence in the Pound)
What are Baxter or BICS Indices?
Baxter or BICS Indices are construction industry specific inflation measures that measure the specific cost pressures of construction specific commodities and services
The plotted trend line of these measures produce the plotted trend line below.
As can be seen the AEI and Baxter Indices have escalated over the period of the data set at a faster rate than the RPIx measure against the notional “basket of goods”. Therefore against the dataset trend lines a trend line needs to plotted that at the end of the dataset this trend line intersects at the high point of whichever dataset has escalated the fastest. The plotted trend line below shows that using the Intellectual Property (IP) of the calculation of the Indexation Model the equation is solved as:
Indexation Factor = RPIx +1.31%.
Why is there IP?
The simple answer is because margins of error have been factored into the calculation of what is effectively the X factor and its method of calculation. Of course this Indexation Model is subject to Copyright as built, but it cannot be subject to a Patent application, as its not an exact science, Its one set of data and a “basket of goods.” To determine its accuracy unfortunately requires a reconciliation at the end of the contract. However its clear at the outset that RPIx + 1.31% is a significantly better level of Indexation for any business that RPI or CPI alone.
This begs the question, “Can the model be adapted?” and the simple answer is Yes, as it’s a dataset of general inflation, a model of industry specific inflation, weighted, and effectively an algorithm to capture the variance. Change any of the input data sets and the algorithm will recalculate itself to give the new required Indexation Factor.
Finally once set is the Indexation Model set in stone for all time?
Again there is no right or wrong answer, but I am currently rebasing the model that I developed and the new data set will run from 1980 – 2014, in effect moving forward by half a notional economic cycle. This will at one level remove the inflationary years where 15% – 20% inflation and wage inflation to match will be removed, but we have to assume that over time the measures self-correct.