It is possible to go around in circles indefinitely looking at competitiveness at the level of its individual components. It is true that there is much more to competitiveness than costs. It is true that Irish labour costs actually compare reasonably well with those of some other regions within the EU. But it is also true that we are already pursuing non-cost aspects of competitivess vigorously, and it’s turning out clearly not to be enough.

There are probably any number of ways to illustrate this, but I have spent a lot of my career looking at policy from a skills and labour market perspective, so I’m drawn to taking a labour market perspective.

As a small open economy, Ireland imports a large share of the goods and services it needs. Over the longer term, these are paid for mainly through the earnings of those working in exporting industries and their subsuppliers and service providers. The value of exports, and even the value added in exports, are misleading indicators because much of the value leaves the country directly through paying for inputs or through repatriation of profits and royalties.

So it is interesting to look at the number of people whose imports each person working in exporting industry pays for. For the sake of simplicity I’ll leave agriculture to one side, as most of the net value it adds to the economy is through transfers from the EU rather than market returns on production.

Data on employment at clients of the industrial development agencies provides a good approximation for employment in exporting industries. This is published annually in the Forfás Employment Survey.

The chart below looks at the ratios between three variables and employment in exporting industries, as measured by the total employment found in this survey.

Ratios to Traded Industry Employment

Over the following 8 years, property-related borrowings from overseas supplemented earnings from exporting industry, allowing the ratio of total employment in the economy to exporting industry employment to rise from 4.9 in 2000 to 6.4 in 2008. Over the same period the ratio of the total 15+ population to exporting industry employment went from 8.6 to 10.7.

With the onset of the current crisis, the flow of net new property related borrowing ceased, causing a severe shock.

My take is that the sustainable ratio between total employment and exporting industry employment for Ireland is probably about the 5:1 that we saw consistently between 1990 and 2000. The policy challenge then facing us is how to restore this balance as fast as possible, and with a minimum of pain.

From my perspective, the only even marginally palatable way forward is to grow exporting employment rapidly. We can’t keep borrowing at the current rate. And if we rely on each person working in exporting industry to bring in enough to pay for their own imports and those of upwards of 10 others aged 15+, that implies quite a drastic cut in living standards. In terms of the numbers aged 15+ each person working in exporting industry is supporting, we are now back to where we were in the early 1990s, still suffering from the aftermath of the disaster that was Irish economic policy for most of the 1980s.

Which brings us back to competitiveness.

While I am sure there is room for improvement, in broad terms we are already doing what we can on non-cost dimensions to improve competitiveness. If we want to significantly increase employment in exporting industries, we have no choice but to address our cost base.

That means tackling:
- Pay
- Returns to all interests in protected sectors
- Returns to all interests in other sectors with weak price competition
- Returns on property

And on pay, those working in exporting industries should only be first in line where there is a direct problem, both because they have lost out on pay since 2000 and because there is no gain to Ireland in cutting pay where an exporting operation is competitive anyway. The deeper problem is in domestically traded and non-market sectors of the economy. See.

The chart below draws on city level data on 46 major US cities which distinguishes between land value and “structure cost” of residential properties. While it refers to buildings that are already in existence, the land value component should give a reasonable degree of guidance as to trends in the the value of undeveloped land suitable for development in those cities.

(Data source: http://www.lincolninst.edu/subcenters/land-values/data/2009-07-DAVIS-PALUMBO.DATA.XLS )

The most interesting thing about the chart is the extreme degree of variation between cities. While the most common level of fall lies in the range 30-40%, values have barely fallen in some cities (Houston and Buffalo), and have fallen by in excess of 80% (Phoenix, Cleveland, Atlanta, Tampa) and even 90% (Minneapolis-St.Paul, Detroit) from peak in others. What that means in the Irish context, is that almost any conceivable level of fall in the land value of property lies within the range of current US experience.

Land value has been much more volatile than structure cost in the US data, presumably because there has been less scope for construction pay rates and the cost of construction materials to change. In this context, it is interesting that Irish construction costs (as measured by the SCS Tender Price Index) are about 24% below peak.

The chart below uses ECB data to track trends in lending to consumers by Irish Monetary Financial Institutions (MFIs), both in nominal terms and discounted by the CPI to track the real value of outstanding lending. The chart runs up to August 2009. I am assuming that the jump in consumer credit numbers between Dec 2008 and Jan 2009 reflects a discontinuity in the data, rather than a real month-on-month jump in excess of €5bn.

While I’m sure it is of little immediate comfort to those depending on selling goods, services and property to Irish consumers, to my mind it is good news that the indebtedness of Irish consumers is clearly falling, in both nominal and real terms. The economy is in a hole largely as a consequence of excessive consumer borrowing, but at least consumers have stopped digging, and are making some gradual progress in refilling their own personal sections of the hole. The real outstanding value of house loans fell by almost 5% between August 2008 and August 2009.

Of considerable practical importance, this means that a significant effective devaluation through reducing price levels within Ireland, while remaining within the common currency, will not necessarily result in an effective balloooning of consumer debt.

Share of All Employment in Ireland Accounted for by State-Dominated Sectors

Data from CSO, QNHS

That is making us more like some much-admired (and not-so-admired) countries some of which are very heavily dependent on employment in these sectors.

Share of Employment in Main State-Dominated Sectors across European Countries, Q1 2009

Of course, we don’t have conscription or big defence forces boosting defence employment, and we don’t have Norwegian oil. And this is coming about mainly through the shrinking of productive sectors of the economy, and high unemployment.

Focusing in on the health and social work sector, it looks from Q1 09 data as if the share of our workforce employed in this area is moving (however involuntarily) into line with high income European countries. The data shown are for Q1 09. Irish Q2 data brings us in line with Q1 German data.

Share of Employment Accounted for by Health and Social Work

The Q2 2009 Quarterly National Household Survey tells an interesting story. Here is chart summarising how the population aged 15+ is occupied, and how this has changed since 2004.

Note that the data used in this chart, and quoted here, are seasonally adjusted, where some QNHS data that you may see elsewhere are not seasonally adjusted.

The sectors that have led the economy downwards, and their employees, have continued to have a very rough time, much as expected by myself, and I would imagine by others making projections.
- Construction lost 23,500 jobs in the quarter. Now at 158,000, or 8.1% of the working population, it can’t keep up this pace of decline much longer.
- Industry (manufacturing, power generation etc.) lost 7,600 jobs.
- Evidence of trouble in employment in agriculture, forestry and fisheries has turned out to be real, with employment falling another 4,300, to bring it below 100,000 for the first time ever.

Most domestically traded market service sectors have lost employment, at roughly the same rate at which total employment fell nationally. This continues a well established pattern, which I have described in papers such as this.
- Employment in wholesale and retail fell by 4,100, with its share of all employment rising just a smidgen, from 14.3% to 14.4%.
- Employment in professional, scientific and technical activities fell by 2,500, maintaining its 5.3% share of all employment.
- Employment in administrative and support service activities fell by 400, maintaining its 3.4% share of employment.
- Employment in Other NACE activities fell by 1,500, maintaining its 5.1% share.

Labour force participation actually rose a shade during the quarter, reversing quite a steep downward trend that had previously been moderating the rate of increase in unemployment.

So with the main drivers of rising unemployment continuing to push forward at full throttle, and with the slight braking imposed by falling labour market participation removed for the moment, what on earth can be causing the rate of increase in the unemployment rate to slow? And why has labour force participation stopped falling, for the moment at least, when it usually falls in tough economic times as jobseekers become discouraged, and where one earner in a two earner family loses their job and decides to ditch the childcare and rely on one income?

Based on the QNHS data, increasing net outward emigration is making a contribution. While the population aged 15+ continued to rise through Q1 2009, it was effectively stalled in Q2, indicating that the full natural increase was being dissipated through net outward migration. But this just meant that population growth was not making matters worse.

More significantly, employment actually increased in a bunch of sectors where one might reasonably have anticipated that it would have fallen.

Part of this effect was concentrated in sectors dominated by public sector employment.
- Employment in human health and social work activities increased by 5,500.
- Employment in public administration and defense increased by 1,000.
- Employment in education fell by 1,400, but the fall was small enough so that education’s share of all employment still rose a little.
The overall share of employment accounted for by these three sectors rose to 24.9%, rocketing up from 24.2% of all employment in Q1 2009, and well up from 20.7% in Q1 2004.

Other sectors in which parallel developments occured were:
- Employment in financial, insurance and real estate services rose by 2,900, in the midst of one of the greatest crises that have ever faced financial services in Ireland, bringing its share of all employment from 5.3% to 5.6%.
- Employment in accomodation and food service rose by 1,400, also in the midst of one of the greatest crises that has ever afflicted the industry, where in the past it has tended to track overall employment in the economy.
- And employment in transport and storage also rose, by 1,100, again in the midst of a great crisis, and where again it has tracked overall employment in the past.

Better job opportunities arising from these developments may be behind the quarterly rise in labour market participation.

I don’t know what went into the models behind other projections of employment heading towards 17%, but mine assumed that employment in domestically traded market services sectors would change (fall under current circumstances) in proportion to total employment in the economy, as it generally has in the past. It also assumed that employment in sectors dominated by the public sector would fall on the basis of constraints on public sector employment announced by the Government. Falling employment in both these areas would have a second round multiplier effect on employment on employment in domestically traded services, driving total employment down even further. Clearly, unless there have been major sampling problems with the Q2 2009 QNHS, these assumptions have turned out to be incorrect.

In searching for an explanation, the common thread I see is that the State has a major influence on employment in all these sectors, whether directly or indirectly. The State is directly responsible for the majority of employment in in human health and social work activities, public administration and defense and education. It must also be having a major impact on employment in financial services, with the domestic banking system being propped up by public funds, with few if any reports of Irish banks moving to bring employment levels in line with their current needs, and with the resulting elevated employment costs arguably being absorbed in the public support for the institutions concerned. In transport and storage, some of the biggest employers are in the public sector.

In accomodation and food service, it is reported widely that many hotels that cannot cover their costs are being kept open, with staff, to satisfy the requirements of tax incentive schemes that financed them. Even if there is no new public subsidy involved, maintenance of the terms of these schemes is a matter of controversial public policy choice.

All in all, while one would clearly wish to look more widely than labour statistics for evidence, the QNHS Q2 2009 data seem to point towards employment funded directly and indirectly by the taxpayer, or effectively mandated by the State, as being central to the slowdown in growth in unemployment. There is some support for this perspective in the Quarterly National Accounts for Q2 2009, which show real spending on public administration and defense rising by 2.2% on the quarter, rather than falling.

Perhaps there is a rather subtle Irish stimulus package in place after all. It will be disappointing if it turns out that the wider evidence of stabilisation in the economy is built only on such a narrow base.

Roll on the eventual release of Q3 data.

addendum 30/9/09
With today’s release of CSO data showing the rise in the standardised unemployment rate to have stalled, I have been reflecting more on the above. As falling employment in construction, compounded by its multiplier effect on the rest of the economy, has been the dominant factor driving unemployment upwards, it seems reasonable to speculate that the rate of decline in construction employment may have moderated significantly in Q3. This could be partly a matter of hitting approaching natural limits (8% of all employment now, 5% to 6% during the ’80s) and/or the result of a boost to public spending on construction.

There is currently considerable controversy over the prices likely to be paid by the National Asset Management Agency (NAMA) for bank loans secured on properties. Calculations of a long term economic value (LTEV)for these loans will contribute to decisions about prices to be paid. And it seems that information on historical trends in property prices will inform the LTEV calculations. If what I have read is correct, these trends are being calculated from data for the period 1970 to 2000, the thirty years argued to be immediately prior to the start of the residential property bubble.

As an aside, I would personally date the start of the residential property bubble to 1998, and omit a substantial part of the almost 70% increase in real property prices between 1997 and 2000 from any assessment of the pre-bubble trend. (Data snagged from FinFacts, originally from Bank of International Settlement.)

So, did the number of households with more than one earner change much over that period? And if so, how likely is it that this historical trend will continue at a similar pace into the future?

A straightforward way to get an insight into this is to look at labour force participation by married women. CSO Census data show the following progression. (Sorry about layout of table.)

Year Participation Intercensal
Rate Change in Rate
1981 17%
1986 22% +5%
1991 34% +12%
1996 50% +16%
2002 46% -4%
2006 53% +7%

Clearly, there was a very substantial increase in the proportion of married women working over the period 1970 to 2000, which will have driven a substantial increase in household incomes independent of changes in individual pay. Real house prices were 153% higher in 2000 than in 1970 (data from FinFacts/BIS as before), and it seems inescapable that the increase in female labour market participation over that time was among the key factors that contributed to driving prices upwards.

Looking to the future, can we expect to see labour market participation by married women increasing at a rate similar to that seen in the 1980s and 1990s? The straightforward answer is no. Setting aside a dip associated with the current recession, there is still an upward trend, but it is slow and shows no sign of accelerating. So the component of the upward trend in house prices between 1970 and 2000 that was associated with increasing labour force participation by married females should logically be substantially excluded from projections of future residential property prices that inform estimates of the Long Term Economic Value of loans to be acquired by NAMA.

But, hey, they’re bright folks. Hopefully they have figured this out already.

Only a small minority of US universities recover as much as 10% of what they spend on research through licensing their intellectual property to existing companies and start-ups, and most of the best known research universities fall below this level. While the income that universities earn through commercialisation gets a lot of policy attention, its true importance lies in the behaviour it incentivises, rather than in the trivial average direct return on investment that it provides at most universities.

So what do we get of economic value from publicly funded university research, other than a financial return to universities on the intellectual property they generate? I would pick out the following ways in which research benefits industry, emphatically not in order of significance.

- A supply of PhD graduates and people with postdoctoral research experience for industry
- Better quality taught university courses, with better quality graduates, because more teaching is by research active academics
- A significant impetus, with resources behind it, to apply research undertaken locally, in start-ups or existing enterprises, even if most research is never commercialised and even if the majority of the commercialisation of local research is not local
- A significant contribution to the capacity to commercialise and/or absorb locally research that has been undertaken elsewhere
- The raw material for a substantial part of the interpersonal and commercial networks that make up a self-catalysing industry cluster (these networks are arguably the main route through which many major research universities create economic value)
- Research services and research cooperation with industry
- Improvements to the functioning of the Irish economy, which are arguably more likely to arise from research into the social sciences (business, application of IT, economics etc.) than from SET disciplines

I have seen arguments that new technologies developed in a country will materially benefit that country as they are adopted. It is an argument that may make sense for a large country, but on the whole I do not believe it works for Ireland. It will always be the case that almost all of the benefit from adopting technologies developed in Ireland will be experienced elsewhere, and that almost all of the new technologies from whose adoption we benefit will have been developed elsewhere. While it is possible to spin early adoption of Irish technologies in Ireland as benefiting adoptors, it is better understood as a strategy to support innovation.

In terms of industry impact, these ways in which research benefits industry should manifest themselves in:
- Attracting more inward investment, both investment that is directly research oriented, and a wider range of projects from businesses that expect to perform better if located within a functioning cluster
- A substantially greater accumulation of start-ups and their spin-offs and descendents over time
- A higher rate of increase in productivity among existing firms, whether Irish or overseas-owned
- Greater resilience and potential for growth and profitability among existing firms, whether Irish or overseas-owned
- Higher average and median pay
- Economic activity and employment in complementary industries
- Economic activity and employment in domestically traded service industries and construction

An issue with assessing the economic impact of start-ups is that even when it is working well it tends to start off slowly, before gathering pace, so it is difficult to distinguish early on between success and failure based just on raw numbers. It is still early days on commercialising publicly funded university research in Ireland. There is also a more general issue with time lags of varying durations between investment and economic impact.

At times, such as the present, when there is a shortage of jobs, clearly the economic return on the creation of a new job will be greater than that at a time of full employment, when there is a greater likelihood of displacing another job.

While many of the above benefits are captured by parties other than the State, there should eventually be a return to the State in higher taxes derived from greater economic activity.

Any assessment of the economic impact of publicly funded university research in Ireland that aspires to accuracy should endeavour to measure the economic impacts I have described, and ideally place values upon them.

p.s. If anyone comes across this post while preparing a response to the economic impact tendering process currently underway, I am interested in talking.

Indexed to Q1 1998, the chart below looks at how real average weekly pay has changed for a range of sectors. In manufacturing it looks separately at industrial and managerial (managers/administrative/technical) workers because the composition of the workforce has changed quite radically over the period, flattering all-employee average numbers on trends in pay for the sector. (I have omitted the clerical worker category to limit data overload.)

I believe that 1998 provides as good a point of reference as any, because it was a time when people were excited about working in exporting industry, but when the fruits of the Celtic Tiger boom had already made good progress in trickling down to traditionally poorly paid services sectors. It is a couple of years before the peak of the Celtic Tiger exporting boom, and coincides with the early stirrings of the construction bubble.  

As can be seen in the chart , average weekly pay in the “computing activity / R&D” (NACE 72-73) sector, mostly made up of software companies and businesses providing other IT-based services, fell a long way behind other sectors, and has only recently returned to where it was in 2000 in real terms.

In manufacturing, average weekly pay for managerial employees (category includes managers, administrative and technical employees) did better, but still fell well behind gains among public sector, construction and distribution workers, and indeed latterly fell significantly behind those in banking/building societies/insurance. Those in “industrial worker” jobs did a bit better.

I don’t have a detailed policy prescription for how we should respond to these facts, but I believe that competitiveness will benefit if there is a realignment of relative pay between sectors, bringing us closer to the relativities that existed in 1998. We need to return to a position where far more of our best people want to work in exporting industries. Improvements in real pay in computing activity, and among managerial workers in manufacturing, over the last year or so are encouraging in this respect, as is the rising interest in science and (non-construction) technology among college applicants. The drive for improved cost competitiveness should be seen in protected sectors of the economy, and among new recruits to exporting industries, ahead of those already working in exporting industry jobs that are viable at current levels of pay.

To those who say that all should suffer equally under our current economic stresses, I would make two points.

1) In relative terms, those working in exporting industries have suffered on pay over the last few years, while others have thrived. It’s time to redistribute the pain.

2) The remuneration of those working in exporting industries is one of the main channels through which the value of exports is captured for the Irish economy, rather than flowing overseas. From the point of view of maximising domestic economic activity and tax take, the greater the total earnings of all those working in exporting industry the better for everyone.

Three final points: 1) It is important to understand that it is significant that the data in the chart relate to full time employees. In industries such as distribution, where there are large numbers of part time employees much worse paid than full-timers, statistics based on the full employee population might look different. 2) I have limited the sectoral coverage of the chart to keep it readable in this blog format. 3) Manufacturing and computer-related services account for the majority of exporting industry. Unfortunately, the time series on financial services pay that includes the internationally traded financial sector has only been running for a couple of years.

The easy-to-measure direct return on public investment in R&D through licensing intellectual property to start-ups and established firms accounts for only a small part of its total economic impact.

Another way to look at the topic is at a more macro level. When one picks through the data, it turns out that there is quite a strong correlation between the R&D intensity of an economy and growth in labour productivity. Here’s a hastily-thrown-together chart based on OECD data that I think demonstrates this adequately. I’ve left countries playing serious labour productivity catch-up (central European, Turkey, Greece, Mexico etc.) out of the chart as they confuse the picture.

Arguably, Ireland, Iceland and Korea were still playing productivity catch-up to some extent over this period, and if they are excluded the data are clustered more tightly around a line.