Wednesday, September 23, 2015

Excavating Irish Archaeology in the 21st century – an imperfect portrait


In February 2014 I published a rather lengthy piece on the recent financial history of commercial archaeology in Northern Ireland. The paper took each of the major companies and proceeded to look at their highs and lows over the period for which ‘Key Financials’ were publicly available. It then sought to average the data to provide an overall impression of the ‘health’ of the commercial archaeology scene here. For the most part, it illustrated a pattern of decay from highs up to 2008, and plunging Net Worth in the years following. At that point, only one company (incorporated in 2005) had posted results for 2013, but it was sufficient to suggest a modest recovery across the sector. By the time the 2013 results for two further companies were made public, it was clear that this one bonanza year couldn’t mask the poor performance of the other two outfits. While there was still an overall uptick in results, it was nowhere near as pronounced as it had previously seemed.

With the publication of both of these posts, I received a variety of responses along one central theme: Norn Iron’s all well and good, but it’d be great to see what the situation is like in the Republic. My thoughts at the time were quite simple – it’s a much larger proposition trying to conduct the same sort of survey for the south … many more companies, more difficult to access the data, and much more complex financial histories … and that was just the start! Beyond that, I’m afraid that my energies just didn’t stretch that far … most of my excavation career was spent in Northern Ireland … this is where my interests lay and where I keep the axes that I still grind. But it was like a nagging itch that wouldn’t go away … a scab that I couldn’t help picking about the edges. Even an exploratory engagement with the data indicated that the format I’d used for the NI companies just wouldn’t work here. There were simply too many companies to go through in any depth and creating a graph and a data table for each was simply impractical. I needed a different approach. As regular readers of this blog may know, I’ve been trying to leverage some of the functionality of the Tableau business intelligence application to present historical data. I’ve used it to look at the spread of early printing presses, the battles of World War I, the battles of the US Civil War, and viewer figures for Game of Thrones (because, why not?). After all this time, it seemed almost counterintuitive to use Tableau to do what it’s meant to do – business!

The finished Dashboard with data from 2001-2014
The Data
First of all, let’s look at the data and how it’s structured, what it is, and how it was collected. I started with a simple search on the Goldenpages.ie website for archaeologists [here]. I searched the CompanyCheck website for companies with ‘Archaeology’ in their title, selecting anything with an Irish address and submitted accounts. I searched for anything I knew should be on the list, but wasn’t. Finally, I was given additional data by a collaborator who has been pursuing research along a similar line to my own … that’s as scientific as it gets, folks! Taken together, this gave me a ‘longlist’ of 159 companies of one kind or another, including sole traders advertising their services. When whittled down to just those companies that have submitted accounts, this list is reduced to 38 companies of various sizes. I removed two further companies that appear on the Goldenpages list, one is a Northern Irish company and the other is GB based. While I would have liked to include their shares of the Irish market in the dataset, there is no way of telling how much of their revenue was generated in the Republic, so it was the pragmatic choice to exclude them. For the remaining companies I used data publicly available from the CompanyCheck website that is, in turn, taken from year-end submissions made to Companies House. The strengths of these data are that they are simple, robust indices of financial health that are relatively uniform from enterprise to enterprise and do not involve having to wade through the frequently labyrinthine accounts themselves. What makes this data difficult to work with is the relatively limited nature of what’s available. The records for the majority of companies still in business may be frequently mined for data across four or five years. However, this gives little indication of what their performance was like in the period beyond this – and some have (or had) been going since the mid and late 1990s. Where data survives from earlier times, it invariably comes from business that have ceased trading and been dissolved. Inevitably, this skews the data to a significant extent. In some respects, this is reminiscent of the survival rates of archaeological data, where older and older material becomes harder and harder to find as it has statistically less chances of survival. Rather than attempting to hide this deficiency, or abandon the project entirely, I want to highlight it as a caveat. For this reason, I’ve included a histogram called ‘Data Points’ on the Dashboard. When data from all of the companies are viewed together it is clear that the numbers of available accounts only really begin to become robust around 2009 (with 16 instances) and peak in 2010 with 26 sets of data. Before 2009 the numbers of Key Financial data points are significantly smaller, reaching a peak of just nine in 2005 and being totally absent for 2003. This just leaves the ‘what’ of the data. What have I got that I’m using to tell this story. These Key Financials are in four parts. The ‘Cash at Bank’ is what it says – the liquid cash at the bank along with cash in hand. ‘Net Worth’ or 'Book Value' is the amount of the Shareholders Funds, less the Intangible Assets. The 'Current Liabilities' data is the total sum owed to creditors, bank loans, overdrafts, and all other short-term financial liabilities. The final data point, 'Current Assets', is the total of stocks, debtors, cash, and any other miscellaneous assets due within one year. It is important to note that this latter figure will occasionally differ between data supplied by CompanyCheck and other providers, as not everyone includes the cash balance as part of this figure. The last time I undertook this type of analysis (October 2014), I also used data supplied by another provider that has since become defunct. I’m sorry that this source is no longer available as it provided an interesting breakdown of the Current Assets figures and clearly showed how much of these assets were in the form of dangerously unreliable monies owed by debtors – hardly much grounds for comfort! At that time, the Northern Irish data indicated that Current Assets were made up of between 43.30% and 95.07% monies owed by debtors. While this seems quite variable, the average for the four NI companies was 73.43% being in the form of debtors. While I do not have the same level of detail for the Irish data, we can safely presume that it’s not much different.

When I first started looking at this form of data, my instinct was to make it completely anonymous. My interest was primarily in telling the overview story, rather than dwelling on the fates of individual enterprises, however much fun that may be. However, the legal advice I received at that time was that I should be able to prove, if challenged, the source of my data and any conclusions drawn from it. For that reason, I provided an appendix at a slight remove from the text that linked the pseudonym (in that case, the year of incorporation) to the CompanyCheck records I’ve used. While I’ve not used the names of the archaeological companies directly, they could be found. As several companies in the current dataset were incorporated in the same years, I’ve had to come up with some different means of differentiating them. Based on the number of companies that, certainly in the past, were identified by three initials, I used to joke that, should I ever throw caution to the wind and start my own archaeological company, I’d call it TLA … Three Letter Acronym. With this in mind, I’ve assigned each company a randomly-generated TLA of its own that does not correspond to its real name. Anyone wishing to find out the ‘who’s who’ behind the data can find their answers in the appendix I have provided [here]. In this way, I can uniquely group the data while not directly revealing the individual identities behind it, and still satisfy the requirement to allow accessibility for error checking and rebuttal.

The Dashboard
The main image on the dashboard is a line graph that attempts to provide summed data for the entirety of commercial archaeology in the Republic of Ireland from 2001-2014. It is given in absolute values, adding together all the positive and negative results for each of the Key Financial variables for each listed company. As such, it seeks to present a high-level wholistic view of the entirety of the southern Irish commercial archaeological sector. Hovering over any point on this image brings out a ‘tooltip’, giving the year, the name of the variable, and its value. Below this, to the left, is a reworking of the upper graph as averages for each of the Key Financials. It seeks to give an idea of an hypothetical ‘average’ company working in the Irish market over this period. Below this again is the Data Points histogram, discussed above, giving details on the numbers of company accounts making up each Key Financials variable. When the user hovers over any of the histogram bars a simple ‘tooltip’ gives both the year and the number of records. In the bottom right quadrant is a data table giving all figures for the selected data. Although it’s pretty redundant, I’ve added reactive grand totals for each column ... mostly because Tableau allows me to do so and I like it.

As is usual for the Tableau environment, the right-hand edge contains the user-defined controls. At the top the ‘Year’ slider allows the user to select a continuous run of years, trimming off data points from either end, to provide a view of the data that’s of most interest and relevance to them. This acts on both of the line graphs, the histogram, and the data table. Below this is the ‘Key Financials’ control that allows the user to select all or just some of the categories outlined above: Cash at Bank, Net Worth, Current Liabilities, and Current Assets. Owing to technical considerations, this only acts on the main line graph (Absolute Values) and not on the Average Values one. Next comes the 'Active ?' control, indicating whether or not the companies in question are still operational. Below this again is the list of companies, coded by their randomly assigned ‘TLA’.  The colour key to the Key Finanacials variables showing ‘which is which’ is located in the bottom right corner of the dashboard. Taken in combination, the user can easily and seamlessly move from the top level view of all companies, over all available years, and all Key Financial variables, right down to a single variable for one company in an individual year … or anywhere in between. Obviously, if a single company is selected, the main line graph and the Average Values graphs will be identical!

An Imperfect Portrait
Given all these caveats, we can be clear that there are myriad ‘known unknowns’ and we can be quite sure there are plenty of the dreaded ‘unknown unknowns’ lurking out there too. But taken at face-value, what story does this data tell? In the first instance, the Cash Story shows a considerable rise to something in excess of €1.8M in 2004, dropping to almost €1.5M the following year. This then continues to rise, eventually peaking at around €6.8M in 2008, only to fall to just over €1.97M in 2009. This 2008 to 2009 period is, obviously, the critical point in terms of the effects of the economic downturn. Personally, I was surprised to such a rapid recovery in the time following, with rises to €4.75M in 2010, and €6.52M in 2011. These figures decline to €3.5M in 2012 and €2.9M in 2013. Given the relatively small numbers of accounts currently available for 2014 (11 sets of accounts, vs. 19 for 2013), the apparent catastrophic drop to c. €1.86M is not a reality, merely a matter of lack of data. The Average Values graph provides a valuable calibrating influence here, showing the same massive drop from 2008 to 2009 and a much smoother, softer recovery than the main graph. The Average Values graph also shows a marked dip in 2005 that does not appear to be related to a smaller number of available accounts. It would appear that this dip is ‘real’ in the sense that it accurately reflects the available data. However, it would appear to be an artefact of a larger number of relatively poor sets of accounts. While there may be a genuine set of economic reasons and events at work here, it would be best to treat data of this age with caution. One of my collaborators suggests that the 2005 dip reflects 'the period of the dotcom crash and the beginning of the second phase of the celtic tiger fuelled by revenues from construction rather than the export driven first phase between 1998-2004'. The story of the Current Assets data is broadly parallel to the Cash story throughout this period. There’s continuous upsurge through the available data, to an all-time peak of €16.8M in 2008, with a fall to 9.58M in 2009, followed by recovery into 2011 and decline through 2012 and 2013. The Average Values data is pretty similar, showing the 2005 dip, the 2009 crash and a steady tail off in the years since, with no obvious signs of growth or recovery. Differences between the two images would include the Average Values don't show an immediate bounce back in 2010, thought there is continued decline from 2011 onwards. Also, the Absolute Values show a marked peak in 2008, while the Average Values paint the picture as more of a 2007-2008 plateau. Overall figures for Net Worth, not unsurprisingly, closely mirror the Current Assets data for both Absolute and Average values. Here we see a relatively gentle rises to c.€2.9M in 2005, reaching its zenith of c.€16.6M in 2008 … all before falling to €7.8M in 2009. Think about it … the best part of €9M was wiped off the value of Irish archaeology as a commercial concern in just one year … that’s a loss of almost 50% of the value of the entire market. Worse than that, the combined Net Worth fell further to c.€6.25M in 2010. Despite making some of this ground back in 2011 (€6.35M), there has been a steady downward decline to c.€2.19M in 2013. The situation is largely mirrored in the Average Values data, showing the ‘2005 dip’, the 2007-2008 plateau at c.€2M, dropping to €487K in 2009, and down to c.€424K in 2010, and only minor changes and fluctuations in the time since, showing general decline. Current Liabilities tell a different and, frankly, more disturbing story. From remarkably modest levels of debt at the beginning of the century, liabilities increased inexorably over the following decade. By the peak in 2008, liabilities had increased (in absolute terms) to c.€3.28M. However, by the time the other variables crashed in 2009, Current Liabilities had increased to €4.81M. This is understandable, as although profits and Net Worth etc. may have crashed, company debts would not have reduced in the same manner. While the other variables largely indicate a partial recovery by the following year, the Current Liabilities increased to a massive c.€6.46M. The story in the time since has would appear to be one of steady debt reduction, through whatever means. The Average Values graph tells a somewhat different story. Here, Current Liabilities generally increased from the beginning of the decade up to 2008, where they hit €410K. The average liabilities, mirroring the other Key Financials, shrunk in 2009 to c.€300K. They remained relatively steady from this until 2011, and in decline in the period since, hitting €124K in 2013.

Well … that’s what the data appears to show … and that’s why it’s an imperfect portrait. No more than ‘normal’ archaeological data, these used here include caveats and lacunae. For all the issues that surround them, it is clear that they do give an overall impression of this period that is consistent with the accepted narrative of growth in the sector up until 2008, followed by a catastrophic crash and some combination of recovery and stagnation in the period since. While the available data undoubtedly capture a good likeness of the crash and its aftermath, the data from before this point is sadly lacking. For this reason, I think that although this exercise has some merit – chiefly in attempting to put actual Euro values on the entirety of the sector and attempting to quantify the extent of the crash – there remains much work to be done to expand and flesh out this portrait of an industry.


The dashboard with data selected only for still-active companies
My original blog post ended there. However, one of my collaborators suggested that I refine the data further to allow the exclusion of companies no longer in existence. In this way, a secondary portrait could be constructed - one of the survivors of the financial crisis who have gone on to dominate what's left of the archaeological sector in the post-Celtic Tiger landscape. These are the companies still in business, still excavating sites, and still providing employment. Screening out all the inactive companies, and the ones where there has been a significant time lapse since the last set of published accounts (the 'Unknowns'), we are left with some 18 companies that have associated data covering the period from 2009 to 2014. In some respects, it is unfortunate that the data doesn't go back even a single year further as it would have neatly captured the 'prelapsarian' state of the archaeological economy. However, this 'After the Gold Rush' information is valuable as it clearly shows the attempts at recovery from the aftermath. From this we can see that the story told by the Cash at Bank show a recovery to c.€2.8M in 2010, from a low of c.€848K in 2009. This increased to c.€3.6M in 2011, but has been in decline ever since, hitting €2.9M in 2013. The Average Values graph tells a similar story of increase from a low in 2009 (c.€106K) to a high in 2011 (c.€212K), and decreases in the years since to c.€173K in 2013. The Absolute Values of Current Assets tell a broadly similar story of increases to €6.36M in 2010 from €3.32M in the previous year, and continuing decline in the period since, down to c.€4.47M in 2013. The Average Values for this Key Financial, however, tell a quite different story, one of continued decline from €415K in 2009 to €262K in 2013. The Average Values for Net Worth tell a similar, if parallel, story of steady decline from €260K in 2009 to c.€129K in 2013. In both cases, there is a noticeable uptick for 2014, though this would appear to be an artefact of a smaller number of returns submitted in 2014 (10 sets) as opposed to 2013 (17 sets). It is for the same reason that a degree of caution should be exercised in comparing the 2009 to 2010 results (eight sets of accounts vs. 17 sets). In terms of Absolute Values, the Net Worth metric may be seen to rise from an historic low of c.€2.07M in 2009 to a high of c.€3.13M in the following year, with gradual decline in the period since, to some €2.19M in 2013. The story told by the Current Liabilities data is very similar to the overall view in that it hits c.€1.58M in 2009, but continues to increase in 2010 to €3.88M, declining ever since, eventually hitting €2.35M in 2013. The Average Values data is broadly similar, showing Liabilities at €198K in 2009, increasing to €228K in 2010, but continuously falling, hitting €138K in 2013. One of my collaborators in this project has suggested that this reduction in the overall levels of debt could, in part, reflect companies divesting themselves of expensive fleets of leased motor vehicles, and surrendering leases on office and warehouse space. If this is the case, it raises new and troubling questions about the storage of the physical and digital archives from previous excavations.

Looking at the individual surviving companies - even on the limited timescale available - it is clear that all have suffered some form of financial pain. For some it was in 2010, or 2013, rather than 2009. For some there have been rapid rises and falls, while for others it has been a more steady, gradual decline. Some have shown a remarkably resilience, bouncing back to having their best years on record. For some, the most recent set of accounts show their worst financial years yet. While there are a number of Lamontian 'green shoots' visible, the general story is one of continued decay and stagnation. It is tempting to see the 2008-2009 period as nothing less than an extinction-level event for the Irish commercial archaeological sector. While several of the mega-fauna trundled off into extinction around this time, there has been relatively little evidence of others striving and adapting to fill the recently-voided niche. There's a reason for this ... there's still one of the great predators stomping about the jungle. For example, in 2012, the combined Current Assets of the other 16 companies came in at  €2,838,655, while for this one enterprise it was €2,329,920. To put that in context: this one company represents 45% of the total amount. And it's not just this one Key Financial either. In the same year, the Net Worth of this company was €1,774,897, versus €992,845 for the others, combined - 64% of the total. Their Cash at Bank was €1,928,488 (54%) as opposed to €1,594,111 for the rest. Significantly, their Current Liabilities were €673,683 (26%) in comparison with €1,880,273 for the combined remainder of the field. By contrast, 2014 was their worst year on record. Even still, they registered Current Assets of €1,600,066 against €1,431,318 for the combined other eight companies for whom data is available. This figure represents 53% of the total. Similarly, their Net Worth of €1,563,583 (66%) compares favourably against the combined €782,971 for the rest. The story is the same for the Cash at Bank, with this one company holding €971,251 (52%) against €897,482. The only real change is that this company's Current Liabilities rose to €382,841 (36%) against some €682,148 for the others. In short, this one company is the most cash rich, is worth more, has more assets, and fewer debts than the rest of the sector combined.

But what does this mean for the future state of commercial archaeology in the Republic of Ireland? Right now, I think it's too early to say. Despite there being some signs of financial recovery, it is extremely unlikely that the sector will see anything like the period around 2007-8 ... at least not in the short to medium term. For now, it seems that the sector is in, if not actual decline, then a period of prolonged stagnation. Only time and additional data will tell. In the meantime, I'll keep collecting data and attempting to paint a better, more complete portrait. In any case, the data and my Dashboard are available to all with an interest, to drill into and find your own stories, present your own portraits, or examine the extinct and future mega-fauna of the Irish archaeological landscape.




Notes
For the best viewing experience of the Tableau dashboard, I would recommend going to Full Screen mode (F11) … there will be less scrolling needed!

The CompanyCheck website gives all data as pounds (£), but it is clear from the more detailed accounts that this is merely a labelling error for the Irish companies, and all are meant to be shown in Euros (€). The obvious exceptions to this should be company accounts for 2001 (as the RoI changed from the Punt to the Euro on January 1 2002). However, of the two sets of accounts currently in the dataset, one is clearly labelled as being calculated in Euros (SRJ) and the other (GSI) does not have the currency listed. In the face of this, I have elected to leave this one example as is and presume that it is calculated in Euros.

In so far as I can ascertain, the CompanyCheck data is correct and without error. The one exception to this is the entry for one company that lists two different sets of data for the year 2010. In the absence of any other evidence, I have assigned the earlier data set to 2009.

Access the dashboard directly at the Tableau server here.