The potential impact of a removal of accountants from the Australian Government’s medium to long-term skilled occupations list
The principal authors of this report are Professor Glenn Withers, Professor of Economics at the Australian National University and Dr Janine Dixon, Senior Research Fellow at the Centre of Policy Studies at Victoria University.
Report management was also provided by Professor Peter Abelson, Managing Director of Applied Economics Pty Ltd.
This report provides results from modelling the potential labour market and broader economic impacts in the event of removal of accountants from the medium to long-term skilled occupations list (MLTSSL).
The report first provides a brief review of the labour market for accountants and the nature of skilled migration for accountants, then moves to model the impacts on the economy of removal of accountants from the MLTSSL. Comment is then provided on the modelling analysis and its significance.
The modelling finds impacts for accounting in terms of:
● a loss of accounting output
● a rise in accounting prices
● a decrease of exports of accounting services
● an increase in accounting import services
For the wider economy there is:
● a reduction of GDP
● a reduction of GDP per capita
The size of these effects is shown to vary with the response to the ongoing reduction in MLTSSL entry. The more the alternative methods of accounting supply operate, the less are the impacts on accountancy. This is shown by having two scenarios where the MLTSSL accountancy cuts are made and compared with a baseline without any MLTSSL changes for accountancy.
The cut assumed for the scenarios applies to migrants entering under the ANZSCO categories 221 and 132 for accountants. In this study, this is projected to to total 6547 in 2018 (up from 6108 in 2016) if the MLTSSL list does not change for accountants.
The two scenarios are:
● Scenario One where the removal is gradually responded to by domestic accounting supply (by transfer from other occupations plus new graduates) and by the use of other migrant pathways for some accountants. Over three years, supply is restored here.
● Scenario Two which maintains an ongoing reduction in accountants equal to the migration loss, with no new recovery or response to the MLTSSL cut in accounting numbers, though total population remains as for the base case with non-accounting migration increasing to fill the gap.
The scenarios are modelled for this report using the Victoria University Employment Forecasting (VUEF) model which compares the scenarios with a baseline where accountants remain on the MLTSSL.
Chart A shows the impact found for accounting price and output across the range of response represented through the two scenarios.
Chart A. Projected impacts of MLTSSL changes on accountancy: price and output effects, scenarios 1 and 2, 2017-2024
In both scenarios incumbent accountants will gain some wage benefits. But this also represents a major cost of service increase, and wage increase pressure on close substiute managers and professionals in Scenario One It also represents reduced service output in accounting, and similar pressure elsewhere, as labour moves raise costs in source occupations.
These cost rises plus reduced per capita income overall with reduced efficiency (and reduced population for Scenario One) mean further that wages in trades and labour areas are adversely affected, including especially semi-skilled and labourers as these are least able to take advantage of job openings created by the reduction in supply of accountants.
For the economy as a whole, the loss of MLTSSL migrant accountants will reduce GDP and GDP per capita. Accountants are relatively high productivity workers and the policy change reduces their proportion in the work-force. Hence a fall in GDP per capita. And population too can fall if there is no replacement through other migration, also affecting GDP through scale effects in this Scenario One case. GDP per capita also falls if the participation rate falls.
Chart B summarises these effects:
Chart B. Comparison of GDP and GDP per capita results, scenarios 1 and 2, 2017-2024
In numbers, removal of some 6000 accountants from MLTSSL each year from 2017 to 2024 leads by 2024, to an estimated loss of GDP in the range of 0.08 - 0.12 per cent and of GDP per capita of 0.05 -0.08 per cent compared to the baseline where the intake of accountants is maintained in the MLTSSL.
With an inflexible scenario on the labour supply side contrasted with a three-year adjustment track alternative scenario in this modelling, the outcome is likely to be within these bounds. That is, because of this modest policy change regarding around 6000 accountants in migration entry restriction, annual real ($2016) GDP would be $1.5- $2.5 billion less than it otherwise would have been in 202
Applied Economics Pty Ltd recently analysed the labour market conditions for accountants, the role played by skilled migrants in addressing current and emerging labour market shortages and gaps, and the international education impacts of the same (Applied Economics, November 2016)
This analysis concluded that there were major ongoing shortages in supply of accountants, especially experienced accountants, and that immigration of accountants for Australia was a crucial element in ensuring an adequate supply of these skills essential for national development.
It was evident that a danger to this contribution by accountants was the discrete ‘on/off ’ nature of the current listing process used for assessing much immigration of skilled migration applicants. Were a delisting to be considered for accountants, albeit mistaken, under the present scheme, there is concern at how much Australian jobs and growth could be compromised. An alternative ‘sliding point’ scale for occupations was recommended as a desirable reform for selection.
This new report has been commissioned for Applied Economics to work with Victoria University’s Centre of Policy Studies (CoPS) to model these potential labour market and broader economic impacts in the event of removal of accountants from the medium to long-term skilled occupations list (MLTSSL), documenting the potential effects in detail.
This report first provides brief labour market background for accountants and the nature of skilled migration for accountants, then moves to model the impacts on the economy of removal of accountants from the MLTSSL list for skilled migration. The accounting migrant loss persists for the economy for every year.
The modelling uses the ‘state of the art’ Victoria University Employment Forecasting (VUEF) model and compares simulations for when accountants remain on the MLTSSL, and where accountants are removed from the MLTSSL. Some commentary is provided on the analysis and its significance.
In the recent Applied Economics analysis of the labour market for accountants it was found that, in the short run there were evident shortages of accountants in specific cases and for specific skills especially for accountants with more experience (Applied Economics, November 2016). These short-term shortages were seen by the study as best being dealt with through short-term visa arrangements. It is therefore appropriate that proposed visa changes announced in May 2017 and as amended in June 2017, retained accountants on the short-term skilled occupations listing (STSOL) too for such visas.
At the same time, in the medium term of 10 or so years, total openings for accountants in accountant jobs of around 11,000 per annum appear likely. In addition there is significant but unquantified demand for persons with accounting qualifications and skills working in management or related financial and other services.
Turning to domestic supply, domestic completions of accounting Bachelor degrees are about 2,500 per annum and not increasing. Domestic completions of accounting majors in all degrees are averaging about 4,700 per annum. But many of these graduates choose careers other than accounting.
Traditionally, immigrants with accounting skills have numbered about 8,000 per annum, including some who do not seek work formally as accountants. This varies over time. Through the early 2000’s, migration of accountants under the skilled permanent entry migration programme increased steadily, and after a brief downturn during the global financial crisis, peaked at over 15,000 migrants in 2010-11. Australia’s intake of migrant accountants has since remained in the range of 6,500 to 8,000 per annum (Figure 1).
The large gap between the forecast openings for accountants as accountants and other demands for accountants, and domestic supply has to be filled by international students staying in Australia and/or by new migrants. The issues are particularly acute in, though not at all limited to, regional areas. Most of those with accounting qualifications find employment relatively quickly in jobs commensurate with their skills.
Currently there are over 39,000 international students enrolled in degrees with accounting majors (Bachelor and Masters). They contribute over $1.7 billion to the Australian economy of which almost $1 billion is Australian universities’ fee income (Applied Economics, November 2016).
Additional benefits include the positive contribution made directly to the economy by overseas-born but Australian-trained accountants who choose to migrate to Australia. If significant changes were to be made to eligibility of foreign accountants for migration to Australia this would certainly have major impacts on universities.
The further impacts on the wider economy of such changes is the subject of this present report, where simulation modelling of removal of accountants from the MLTSSL skilled migration list is provided. With exclusion, a prospect of significant short-term and long-term shortage is found.
To undertake this modelling a well-attested economic model is used. This is the Vic Uni (VU) model which is the core structure used for the analysis. This is the descendent of the well-known Monash model and is thoroughly documented (Dixon and Rimmer 2002). It is a so-called ‘computable general equilibrium’ (CGE) model allowing simulations of the impact of major economic changes, including policy changes, and provides considerable detail of these impacts through-out the economy.
CGE modelling in Australia is at a best practice frontier globally and Australia was one of the early leaders in this method of analysis. Such models allow for the detailed interaction of effects of changes imposed on the model and can provide considerable examination of the component implications. This modelling is widely used for informing public policy in Australia.
Importantly for the CPA Australia exercise, the Victoria University Employment Forecasting (VUEF) model extends the VU model by adding significant detail on occupation and skills. Building on the standard CGE framework in which the demand for labour is described through industry production relationships, VUEF also links the supply of labour by occupation to the growth rates of the workforce classified by skill including experience level (J.Dixon 2016; J.Dixon and Wittwer 2015). Details are at the Appendix to this report.
This approach makes explicit the supply side of the labour market, as well as demand. The VUEF extension is integrated into the Vic-Uni model, thereby enabling feedback within the model from the consequences of labour supply shortages or surpluses. Although markets are assumed to clear, so that shortages and surpluses never eventuate, potential supply shortages shortages or surpluses are indicated by wage pressures in the model.. Importantly for this report, the model represents occupational supply constraints at a fine level of detail.
The Vic-Uni model with the VUEF extension produces annual forecasts over an eight-year horizon of employment cross-classified by industry (115 input-output industries) and occupation (97 ANZSCO minor groups), and by skill (56 groups) and occupation
The modelling results are provided from a “what-if” analysis – that is, the results from the scenarios used here are contingent on the assumed changes in the underlying composition of the workforce. This detail on the balancing of supply and demand is highly informative for policy.
In this report, we present modelling scenarios for policy change which remove all present skilled permanent migrants entering under the MLTSSL arrangements as Accountants. We examine the potential impacts of the removal of accountants from the MLTSSL.
This is presented as a deviation from a baseline where there is no change in the treatment of accountancy in the MLTSSL. This base is an extrapolation flowing from direct consideration of actual trend forces.
This baseline can then be compared with new policy scenarios that arise by instead assuming there is now no MLTSSL accountancy listing. The policy scenarios differ here according to what is assumed about responses, but both remove MLTSSL accountancy listing.
In illustrative scenario 1, the removal of accountants from the MLTSSL is not explicitly offset by an increase in other migration, so the workforce declines in size. However, in subsequent years, the supply of accountants is assumed to recover, through a combination of migration under alternative visas and greater supply from the domestic population. As such, growth in the population and in the supply of accountants both recover, with the supply of accountants recovering slightly more quickly.
In illustrative scenario 2, the removal of accountants from the MLTSSL manifests as a permanent reduction in the growth rate of accountants. The growth rate of the population is assumed to be unaffected, meaning that while total immigration is unchanged, the composition of immigration is permanently altered. There is no response though in the relevant time period through domestic education supply or alternative migration pathways.
These policy change scenarios are further described below. For the baseline itself, a standard base case scenario is projected for the VUEF model. This represents the situation where accountants remain on the MLTSSL. In this case model forecasts are generated by solving the model in response to likely structural change, updating the database from one financial year to the next through a series of simulations.
“Shocks” to the model so as to provide for a common baseline going forward constitute the assumptions made for the baseline case scenario and are based on a combination of:
1. calibration to forecasts from expert bodies, including:
● Bureau of Resources, Energy and Environment for forecasts on value and volume of commodity production and exports;
● a combination of ABS and state demographer forecasts for state population forecasts;
● Intergeneration Report forecasts of participation rate;
● Commonwealth forecasts from the federal budget on taxation and government expenditure;
● Commonwealth forecasts from the federal budget on other key measures including the terms of trade and the unemployment rate;
2. an assumed rate of productivity growth based on recent experience in Australia and similar countries;
3. significant announced changes in economic activity, for example, the closure of motor vehicle manufacturing plants in Victoria and South Australia;
4. likely changes in construction activity revealed in building approval data; and
5. the continuation of structural trends revealed in the database calibration process.
This projection therefore will reflect likely economic drivers and their resultant demand and occupational supply trends, with these being derived for accountants including through MLTSSL- based migration specification being included.
Given the baseline, to examine a policy change on accountancy migration, further shocks are then applied to the model to represent the alternative hypothetical scenario in which accountants (ANZSCO groups 221 and 132) are removed from the skilled occupation list from 2018. Otherwise here the same baseline assumptions, indicated above, continue to apply, except for these specified changes in MLTSSL listing for accountants and changes that flow through the model’s representation of adjustments to this new policy.
For understanding the effects of the new MLTSSL policy setting as compared to “all other business as usual” as embedded in the baseline, it should be recognised that the accounting profession includes two related but different groups of people. These are people qualified and working as accountants and people with accounting qualifications (e.g. majors in accounting in mixed degrees) using their skills other than as an accountant. The qualification definition is especially pertinent in examining economic consequences of reduction.
For this study, individuals qualified as accountants are defined as those with a level of skill commensurate with an Australian Qualifications Framework (AQF) Level 7 or higher qualification (AQF 2009) who nominate their highest qualification as being in the field of accounting.
By contrast, those working as accountants are defined as those working in occupations defined within the Australia and New Zealand Standard Classification of Occupations (ANZSCO, 2006) four digit groups, namely:
2210 Accountants, Auditors and Company Secretaries nfd
2212 Auditors, Company Secretaries and Corporate Treasurers
2220 Financial Brokers and Dealers, and Investment Advisers nfd
2221 Financial Brokers
2222 Financial Dealers
2223 Financial Investment Advisers and Managers
The specific new Scenario One extra “shocks” imposed for the new SMTSSL Scenario one in this study to reflect the new policy are:
1. SMTSSL Categories reduced:
● The supply of workers of the skill type “Bachelor Degree – Management”, which encompasses bachelor-level accounting qualifications, is reduced, by 6547 persons in 2018 (representing anticipated migration of accountants in 2018), and by smaller amounts in 2019 and 2020 (assuming that the reduction in migration is gradually replaced by locals and migrants on other programs).
● The supply of workers of the skill type “Bachelor Degree – Management” TO the #-digit ANZSCO occupation 221 Accountants, Auditors and Company Secretaries is reduced by 6401 in 2018. This is accommodated by changing the preferences of this skill type for this occupation. Thereafter, preferences remain at the 2018 settings, and the employment choices of this skill group are made endogenously.
● Similarly, the supply of workers of the skill type “Bachelor Degree – Management” TO the 3-digit ANZSCO occupation, Business Administration Managers is reduced by 146 in 2018.
2. Recovery through domestic supply and alternative migration pathways:
● The total working age population is reduced by 6547 in 2018 and smaller amounts in 2019-2021 (assuming the reduction in migration is gradually replaced by migrants on other programs). Note the working age population “recovers” more slowly than the supply of bachelor degree-management workers because we assume some of the recovery of the “Bachelor degree – management” skill group is supplied from the domestic population.
● Total employment is reduced by the same amount as the working age population, i.e. 6547 in 2018 and smaller amounts thereafter.
3. Consequent Labour Market Adjustments:
● Model closure – with total employment and the working aged population both determined exogenously, the model endogenously determines the unemployment rate. With total employment assumed to follow the growth path exogenously imposed, the model endogenously determines the real wage.
Figure 2 summarises the resultant significant changes assumed in total employment and in working age population for accounting workers.
We describe here the impacts in key dimensions of the economy resulting from the reductions specified in removal of accounting from the MLTSSL. As seen in Figure 2, there is a migration-related reduction in accounting employment of over 12,000 and of the total so qualified of over 15000.
What these reductions do for the accounting sector and the wider economy assuming some domestic supply response and other migration pathway response, is the subject of this section of the report.
A key feature of accounting is that more than 75 per cent of accounting services are into intermediate production, and a further 14 per cent are sold into investment. Accounting services face little competition from international imports. As such, demand for accounting services is very inelastic.
When growth in the supply of labour to accounting services is reduced, growth in accounting output slows. Because demand for accounting services is very inelastic, this leads to accelerated growth in the price of accounting services, which is passed on to consumers of accounting services, as is seen in Figure 3.
In Figure 3, the quantity of accounting output falls slightly, but because of the price increase, there is an increase in revenue in the accounting industry. Also, there is a temporary increase in capital rental returns in accounting but this is eliminated by an increase in investment. Imports of accounting services increase but from a small base (Figure 4).
For any year, cumulative deviation results can be interpreted as the difference between the “policy” scenario and the baseline (what otherwise would have been). For example, in Figure 3 we see that the result for the accounting price index in 2024 is 1.03. This means that in 2024, the accounting price index is 1.03 per cent higher than it otherwise would have been in 2024, as a result of the changes listed above.
Figure 5 shows two measures of employment in the accounting industry. The “wage bill weights” measure calculates employment as an aggregate of all occupations weighted by wages, while the “persons” measure is a simple headcount of employment. The implication of the difference between the measures is that the reduction in employment is biased towards university-qualified accountants whose wage is higher than the industry average. Note that the industry as a whole includes lower-cost activities such as book-keeping.
Figure 5 also shows a small response in the capital stock of the accounting sector. In the VUEF model, capital response always lags by a year. The increase in the rate of return in capital in 2018 stimulates investment in 2018, delivering a larger capital stock in 2019.
Aggregate demand for accounting is inelastic because of the high proportion of sales to intermediate usage. The production structure of the model imposes ‘Leontief technology’, ie inputs in fixed proportions, on intermediate use, meaning that intermediate demand is unresponsive to price. Demand for intermediate inputs is therefore reduced only if industry output is reduced. Demand for domestically produced intermediate inputs may be replaced by imports.
In the case of accounting inputs, imports do increase, but do not comprise a significant proportion of supply. Export demand for accounting services is more responsive to price. Exports comprise only a small proportion of the sales of accounting services, but a reduction in exports of accounting services accounts for most of the decline in accounting output.
The three sources of reduced sales – a smaller domestic market, a reduction in the domestic share of the domestic market, and a reduction in exports – are illustrated in Figure 6 below.
Short run and long run impacts on GDP can be verified by further calculations. In 2018, the supply of workers with a bachelor degree in management is exogenously reduced by 1.15 per cent, to represent the removal of accountants from the MLTSSL. These workers contribute around 6.8 per cent of aggregate wage-bill weighted labour input, and labour contributes around 60 per cent to GDP. By assumption, the capital stock for 2018 is put in place exogenously by investment already undertaken in 2017, before the shock is known. The impact on GDP of removing these workers is therefore approximately 1.15% * 6.8% * 60 % = 0.05%. This simple calculation verifies the modelled result for 2018 (Figure 7).
In the long run, say by 2024, the supply of workers with a bachelor degree in management has been exogenously reduced by a cumulative 1.8 per cent. In the long run, the capital stock is assumed to have had time to adjust so that rates of return are back at their base case levels, which effectively means that the accumulated percentage change in the capital stock is the same as it is for labour. The long run impact on GDP of removing the accountants is therefore 6.8% * 1.8% = 0.12%, which verifies the modelled result for 2024 (Figure 7).
Accountants are highly qualified and earn a wage that is above average, reflecting a high marginal product of labour. The reduction in effective employment is therefore greater in percentage terms than the reduction in employment by headcount. This is true within the accounting industry (Figure 5) but it is also apparent at the national level. As a result, GDP per capita declines, as shown in Figure 8.
The impact on real GDP per capita is one summary measure of the overall impact of the policy. A better measure of the impact on the welfare of the domestic population is the impact on real gross national income (GNI) per capita. In the context of this policy, the main difference is in the price deflator used. (GNI also accounts for foreign remittances, which are assumed to have only a small impact in this analysis).
Figure 9 show that the terms of trade (the ratio of the price of exports to the price of imports) initially declines and then accelerates above its baseline level. This indicates that imports initially become expensive relative to domestic income, which includes income from sales of exports. In other words, the purchasing power of domestic income is reduced. Beyond 2020, when the terms of trade begin to increase, this effect is reversed, and real gross national income per capita increases relative to real GDP per capita.
It remains to understand the drivers of the terms of trade result. The increase in the terms of trade in the long run is the result of the reduction in employment and GDP shown in Figure 7. Assuming Australia has some market power in exporting, a slightly smaller economy means that the Australian currency is able to appreciate slightly. The short run decline in the terms of trade is the result of reduced investment activity in the short run.
As the workforce declines, the capital stock also needs to decline for rates of return on capital to return to their base case levels. This requires a fall in investment in the short run, which is offset by a temporary move towards trade surplus – an increase in exports and decrease in imports. The increase in exports causes the temporary fall in the terms of trade.
Long run changes in the relative wages of the occupations indicate pressures in the labour market. Wages for accountants increase by more than 5 per cent relative to the baseline. Wages for selected other occupations – the most positively and negatively affected – are shown in Figure 10 below. In general, there is a positive impact on the wages of occupations that, like accountants, are supplied by workers with a bachelor degree in management. These occupations also suffer a shortage of supply, as suitably qualified workers are attracted away from these occupations and into accounting. Note that ANZSCO minor group 132 (Business administration managers) includes the unit group 1322 (Finance managers) which is also subject to an exogenous reduction in supply, albeit much smaller than the reduction in accountants.
Declining wages are most apparent in the trades, including plumbers, bricklayers, glaziers and so forth. This is a result of the decline in investment and domestic incomes. Investment in dwellings is particularly negatively affected, with negative consequences for residential construction and employment of the trades occupations.
The impact on employment by occupation is, as expected, most significant for Accountants (Figure 11). However, there is a negative impact on employment in all occupations as a consequence of the reduction in GDP. Apart from Professionals, there is an appreciable impact on Other Professionals, Managers, Technicians and trades workers and Clerical and administrative workers. The negative impact on the trades workers is a consequence of reduced investment, particularly in dwellings, that also drives the fall in wages for these occupations as discussed previously.
Employment of Managers, Other Professionals and Clerical and administrative staff is also reduced because of the overall decline in the accounting industry, and because some suitably qualified workers choose to become accountants instead of managers or clerical workers. Note that the initially imposed reduction in supply of accountants of over 6000 workers is never realised, because workers substitute into accounting from other occupations. The initial decline in employment of accountants is only just over 2,000 workers for this reason.
To test simulation results against some alternative assumptions can be a useful methodology. It allows examination of alternative assumptions, and sensitivity of results to these. For the modelling in this report, we have included a further scenario (“Scenario Two”) which makes different assumptions as to the degree of responsiveness in other labour supply pathways for accountancy responding to the removal of accountants from the MLTSSL list. The MLTSSL cut is the same but responsiveness in the labour market does not follow through domestic supply movements and other migration pathways.
In this further illustrative Scenario (2), the removal of accountants from the SOL manifests as a permanent reduction in the growth rate of accountants, and a continual reduction in the stock of accountants in the workforce, as shown in Figure 12. But the growth rate of the population is assumed to be unaffected, meaning that while total immigration is unchanged, the composition of immigration is permanently altered. This simulation is unlikely to play out in the long run, but the results illustrate the impacts on the accounting industry of a situation in which neither the domestic nor immigrant labour market respond to a growing shortage of accountants.
This may seem less realistic than the major MLTSSL scenario featured above in this report. But it may be argued that given the lags inherent in much study and professional experience for accreditation, and also given uncertainty over whether MLTSSL might be extended to a wider ambit than SOL so limiting substitute pathways, that such a scenario is still feasible. Hence this alternative and its results are presented here for comparison. And perhaps the likely outcome could be somewhere between the two. If this is so, then the two scenarios define the likely bounds of policy impact, and Scenario Two is of genuine interest as well as Scenario One.
This Scenario Two remains still a “what-if” analysis – that is, the results from the scenario are contingent on the assumed changes in the underlying composition of the workforce.
The scenario is still evaluated using the Victoria University Employment Forecasting (VUEF) computable general equilibrium model of the Australian economy with particular detail in the labour market
In this simulation, the working age population is assumed to remain on its baseline trajectory, as are the participation rate and unemployment rate. As aggregate employment is thus unchanged, results for macroeconomic indicators relative to the baseline are driven by compositional effects only. Because accountants are highly skilled relative to the population on average, the economy-wide effective labour input declines, and GDP per capita also declines, as shown in Figure 15.
A brief comparison of GDP results from the two scenarios reveals that the fall in GDP is larger in scenario 1. A key assumption under scenario 1 is that population growth slows somewhat as a consequence of the reduction in migrant accountants, whereas in scenario 2, migrant accountants are assumed to be replaced by other migrants such that the population remains on its baseline growth path. Because the assumptions for population growth differ between the two scenarios, the appropriate indicator for comparison of the scenarios is GDP per capita. In scenario 1, this differs from the GDP result, while in scenario 2, in which population growth is unchanged, percentage deviation results for GDP and GDP per capita are the same.
Figure 14 shows that, until 2020, results for GDP per capita from the two scenarios are very close. From 2021 onwards, the impact in scenario 1 of stabilising accountant numbers with more entrants from the domestic market and via other migration streams starts to show, with GDP per capita stabilising at 0.05 per cent below the baseline. On the other hand, the permanent reduction in the growth rate of accountants assumed in scenario 2 feeds into a continuous decline in GDP per capita as shown in Figure 14.
The fall in GDP however is much larger for scenario 1. It allows population to fall, relative to the baseline as the accounting migration is cut, whereas scenario 2 keeps population as for the baseline.
The reduction in availability of suitable workers curtails output from the accounting industry (Figure 15). As the supply of accountants falls further below the baseline, so does output and employment in the accounting industry. As outlined earlier, demand for accounting services is relatively inelastic, and consequently the price index for accounting grows. Strong growth in the accounting price index is reflected in the accounting wage index and rental price, or gross operating surplus in the accounting industry, as shown in Figure 16.
As for scenario 1, with the reduction in available labour, the accounting industry becomes more capital intensive, increasing its capital stock slightly (Figure 17). Again, the difference between the employment headcount and effective employment measured in wage bill weights reveals a compositional change towards lower-skilled workers in the accounting industry.
Results for the decomposition of the sales of accounting services (Figure 18) also reveals similar results to scenario 1, with the majority of the reduction in accounting services absorbed by the export sector.
The impacts on occupation wages indicate where shortages and surpluses of labour may develop. As for scenario 1, by far the largest positive wage deviation is for accountants, auditors and company secretaries (ANZSCO 221), for which wages are 8 per cent above baseline by 2024. Results for the highest and lowest remaining occupations are shown in Figure 19.
Many of the occupations with a gain in wages are also those which gain the most in scenario 1, including financial brokers and dealers, advertising managers, sales and marketing professionals, and accounting clerks. In scenario 1, the occupations which saw the largest reduction in wages were the trades occupations, which were affected by a reduction in dwelling investment. In scenario 2, the trades occupations again suffer a negative impact on wages, but some health-care related occupations, including medical practitioners, health therapy professionals and nurses, are also among the most negatively impacted.
The negative impact on the wages of health professionals can be attributed to the assumption that the size of the working age population remains on its baseline trajectory. Recall that we assume the supply of bachelor degree-qualified management graduates is reduced and replaced by uniform expansion in all remaining qualification groups. Cohorts of workers with qualifications that can supply occupations such as ANZSCO 221 (accountants, auditors and company secretaries) will supply more workers to this occupation and fewer to other occupations (such as financial brokers and dealers, advertising managers and other occupations shown in the left half of Figure 19).
This puts upward pressure on the wages in these occupations. However, workers with qualifications in unrelated fields, such as health or trades, are not in a position to take advantage of the opportunity to enter accounting-related occupations. As a result, the assumed increase in supply of workers of all qualifications (other than bachelor degree in management) puts downward pressure on the wages of workers in health and the trades.
In Figure 20, the total impact on employment is broken down into the eight ANZSCO major group categories, and accountants. Unlike in scenario 1, where there was a fall in all categories of employment, in this scenario, the net impact on employment is zero by assumption. Other than accountants and managers, employment in all occupations increases.
Employment of accountants falls sharply in the first year of the simulation by assumption. The supply of workers with a bachelor degree in management is reduced by assumption, and this cohort is also assumed to supply fewer workers to the ANZSCO 221. From 2019 onwards, the supply of workers with a bachelor degree in management continues to fall (see Figure 12) but the cohort reduces its supply of workers to all of the occupations which it supplies. As a result, employment of accountants continues to fall, but at a slower rate.
As of 2016, we estimate that slightly over 20 per cent of workers with a bachelor degree in management work in ANZSCO 221. The supply of this cohort is reduced by more than 45,000 persons relative to baseline by 2024 in this simulation. We might expect that if 20 per cent of these workers are accountants, employment of accountants would fall by 9,000 persons relative to baseline, yet the fall is around half this amount. The discrepancy is explained by an increase in supply to ANZSCO 221 from other qualification groups, including management qualifications at the levels of post-graduate and advanced diploma. This suggests that workers that are either over-qualified or under-qualified for accountancy work may be drawn into ANZSCO 221.
While the inflow of permanent migrants under occupational skill listings in Australia represents a small absolute number in any single year, there are important cumulative effects over time and ripple effects through the economy. This is shown, by the modelling in this report, to apply for a single but key profession such as accounting.
If accountant entry via MLTSSL was stopped, compared to a situation where it continued as before, the modelling shows that there would be significant and adverse consequences. This applies across the range represented by the two scenarios of reduced accountancy migration modelled for this report.
For those scenarios, within accounting itself there would be a rise in accounting cost. For incumbents this is welcome, but the reduced availability of accounting services and higher accounting costs for others are significant negatives. The negatives ultimately feed through as reduced GDP and GDP per capita.
The impact for accountancy is diminished if, and when, other residents move across to accounting and domestic accounting graduate supply increases and substitute migrant pathways are allowed to operate. Accounting service exports fall as the costs rise and imports of such services increase.
Some consequent flow-on effects, especially of loss of staff from elsewhere to accountancy, do benefit wages in those recruitment source areas. But the cost rises of accounting and related services are such that the level of activity elsewhere is reduced, just as in accounting itself. With reduced GDP, wages in trades and labour areas are adversely affected, especially for semi-skilled workers and labourers.
The long-run impact is decline for the economy in employment, GDP and capital formation. The major overall indicator, GDP per capita falls under both Scenarios, and falls progressively where there is no other accounting supply response to assist.
To put this in numbers, removal of some 6000 accountants annually from MLTSSL leads by 2024 to an estimated loss of GDP of up to 0.12 per cent and of GDP per capita as much as 0.08 per cent compared to the baseline case (where the intake of accountants is maintained in the MLTSSL list).
The cut of 6547 accountants (ANZSCO categories 221and 132) falls to zero in three years for the economy in scenario one as substitution takes place. This produces over three years an accumulated total loss of 12,547 employed accountants, but with no new further loss thereafter.
For scenario two, the cut remains in place thereafter for the whole projection period. No substitution occurs and there is a continued and accumulating loss of the stock of accountants over time, growing each year when new accountants would otherwise have joined the Australian workforce from overseas.
The precise trajectory will depend on how responsive the alternative accounting supply pathways would be over what time scale, and how flexible the broader labour market and other arrangements would be.
However,with an inflexible scenario on the labour supply side contrasted with a three-year adjustment track alternative scenario in this modelling, the outcome is likely to be within those bounds. That is, because of this modest policy change annual real ($2016) GDP would be $1.5- $2.5 billion less than it otherwise would have been in 2024.
The Commonwealth Department of Employment has recently listed accountants as representing the third largest occupation for projected job growth by 2020 (Figure 21). To remove accountancy from the skilled permanent immigration program occupational listings would seem most ill-advised in terms of this requirement and in terms of the likely adverse economic impacts also revealed by the modelling in this report.
Figure 21: Projected Job Growth by Occupation by 2020: Top Five Occupations
Applied Economics (2016), The Labour Market for Accountants and the Skilled Occupations List, A report to CPA Australia and Chartered Accountants Australia and New Zealand, November
Department of Employment (2016), Australian Jobs 2016- Report, Canberra: Australian Government Department of Employment.
Dixon, J. (2016a), Victoria University Employment Forecasts Information Paper. Available from author on request.
Dixon, J. M. (2016, October). Victoria University Employment Forecasts for Australia, 2016-17 to 2023-24. Subscription product.
Dixon, J. M., & Wittwer, G. (2015). The Labour Module in a dynamic, regional CGE model. CoPS Working Paper G-257 .
Dixon, J. M., Dixon, P. B., Giesecke, J. A., & Rimmer, M. T. (2014). Quantifying "Dog Days". Economic Papers , 33:3 pp. 203-219.
Dixon, P. B., & Rimmer, M. T. (2002). Dynamic General Equilibrium Modelling for Forecasting and Policy: A Practical Guide and Documentation of MONASH. North Holland Publishing Company.
Giesecke, J. A., & Madden, J. R. (2013). Evidence-based regional economic policy analysis: the role of CGE modelling. Cambridge Journal of Regions, Economy and Society , Vol 6, pp 285-301.
Giesecke, J. A., & Madden, J. R. (2013). Regional Computable General Equilibrium Modelling. In P. B. Dixon, & D. W. Jorgenson, Handbook of Computable General Equilibrium Modelling (pp. Vol 1A pp 379-470). Elsevier.
Meagher, G. A., & Pang, F. (2011). Labour Market Forecasting, Reliability and Workforce Development. CoPS/IMPACT Working Paper Number G
At the core of VUEF is the Vic-Uni model, a highly detailed and modern computable general equilibrium (CGE) model of the Australian economy. The Vic-Uni model is a descendant of the well documented MONASH model (Dixon and Rimmer 2002). A recent example of a comprehensive forecasting simulation derived from the Vic-Uni model is J.Dixon et al (2014).
The model features several decision-making agents – industries (more than 100), investors, households, external trading partners, and government – making decisions on the supply and use of commodities (more than 100, corresponding to the industries) and factor inputs, and whether to source commodity purchases from the domestic economy or trading partners. Three blocks of equations describe this process:
1. Agent decision-making equations, including equations describing
a. cost minimisation by producers,
b. utility maximisation by households,
c. purchases by government, generally policy-driven,
d. the allocation of investment funds in response to risk-adjusted rates of return, and
e. external trade.
2. Equations describing the price system.
a. Basic prices are derived from a zero pure profit condition on producers
b. Purchaser prices are derived by adding relevant margins and taxes. Every agent in the model has a separate purchaser price identified for each commodity.
3. Market clearing equations to ensure supply equates to demand for every commodity.
These equations are sufficient to endogenously determine the price and quantity of every produced commodity identified in the model.
For non-produced commodities – factor inputs and imports sourced from outside Australia – these equations are sufficient to explain price or quantity, but not both. For these commodities, we assume:
1. that the national supply of land is fixed, although it may shift between uses,
2. that the national supply of labour is determined through a series of supplementary equations linking it to national population,
3. that the industry-specific supply of capital is determined through additional dynamic linkage equations to previous-period capital stocks and investment, and
4. that foreign currency prices of imports are exogenous to the model.
Finally, a block of macroeconomic accounting identities is included in the model for the reporting of variables such as GDP, the terms of trade, the real exchange rate, and other key macroeconomic indicators.
In addition to the exogenous variables in factor markets (the national population, and the short run supply of capital) the “naturally” exogenous variables of the model are those that are not explained by the equation blocks described. These include variables that:
1. describe changes in production technology,
2. describe changes in household tastes,
3. describe changes in preferences for imported and domestic varieties,
4. describe changes in government policy, including rates of taxation, budget position and the composition of expenditure,
5. describe international trading conditions in terms of shifts in export demand schedules, and foreign currency prices of imports, and
6. describe changes in investor risk premia.
The model is solved in a recursive dynamic system, such that the solution for one year forms the base data for the following year.
As an economy-wide general equilibrium model, individual industries are linked not only through transactions with one another, but also through competition for resources, in particular the national supply of labour.
Data and calibration
To parameterise its equations, the model makes extensive use of the Australian input-output tables. However, as input-output tables are typically out-of-date by the time of publication, the database is regularly calibrated to the most recent available national accounts data on GDP, private and public consumption, investment, international trade, the terms of trade and industry value added, as well as other ABS data including population, participation rates, the unemployment rate, the wage price index, and the CPI.
The calibration process provides not only an updated database. In addition, by forcing the model to reflect observed macroeconomic and industry data in its updates from year to year, the process reveals structural change in the economy. For example, a “natural” model simulation uses labour, capital and productivity growth as inputs when determining GDP. However, if labour, capital and GDP have all been observed, then the model treats productivity growth as an endogenous variable. In this way, calibration of the model reveals recent shifts in structural parameters of economic growth, such as changes in productivity, consumer preferences, savings rates, willingness to invest, and conditions in the world economy.
The Vic-Uni model forecasts are generated by solving the model in response to likely structural change, updating the database from one financial year to the next through a series of simulations. The “shocks” to the model are based on a combination of:
1. calibration to forecasts from expert bodies, including:
a. BREE for forecasts on value and volume of commodity production and exports;
b. a combination of ABS and state demographer forecasts for state population forecasts;
c. intergeneration report forecasts of participation rate;
d. forecasts from the federal budget on taxation and government expenditure;
e. forecasts from the federal budget on other key measures including the terms of trade and the unemployment rate;
2. an assumed rate of productivity growth based on recent experience in Australia and similar countries;
3. significant announced changes in economic activity, for example, the closure of motor vehicle manufacturing plants in Victoria and South Australia;
4. likely changes in construction activity revealed in building approval data; and
5. the continuation of some structural trends revealed in the database calibration process.
Importantly for labour market forecasting, the VUEF model extends the VU model by adding significant detail on occupation and skills (Meagher and Pang 2011, J.Dixon and Wittwer 2016). Building on the standard CGE framework in which the demand for labour is described through industry production functions, VUEF also links the supply of labour by occupation to the growth rates of the workforce classified by skill.
This makes explicit the supply side of the labour market. The VUEF extension is integrated into the Vic-Uni model, thereby enabling feedback within the model from the consequences of labour supply shortages or surpluses.
For any industry, the aggregate demand for labour is specified as a function of total output of the industry (which is in turn a function of demand for the industry’s output), the industry’s capital stocks, changes in productivity, and the cost to the industry of labour relative to other inputs. Having determined aggregate demand for labour, demand by occupation is specified as a function of aggregate demand for labour, changes in the wage of each occupation relative to average, and changes in the relative productivity of the occupations.
For the model to generate changes in the relative wages of occupations, the supply of occupations needs to be constrained. In the VUEF model, this constraint is introduced by partitioning the workforce into skill groups. “Skill” in this context relates to qualifications classified by level and field, for example “Bachelor degree: Health”. The skill cohorts supply labour to the occupations, based on relative wage.
From industries therefore, demand for occupations increases as the relative wage falls, whereas from the workforce, supply of labour to occupations falls as the relative wage falls. The market solution is achieved when demand and supply equate.
Growth in skill groups is estimated as a function of projected existing skill cohorts by age group, and skill acquisition by age group.
This formulation introduces a key rigidity into the modelling framework – that workers have limited capacity to change occupation, therefore demand for labour by occupation cannot necessarily be filled at the current wage rate.
The VUEF extension will also have the capacity to generate forecasts of unemployment by skill, last occupation and last industry. Users familiar with previous versions of VUEF will note that this is a new feature. The basis of the methodology is to assume that in addition to the 97 occupations, each skill cohort can also allocate some labour an “unemployed” category. If the average occupation wage falls, labour is withdrawn from employment, reflecting the downward rigidity of wages. Unemployment by skill group is then projected onto former occupation and industry of employment.
Data and calibration
As with the Vic-Uni model, the VUEF extension is formulated at a very fine level of detail. It identifies 56 skill groups, consisting of 5 levels (the ASCED broad categories for post-school qualifications) multiplied by 11 fields (the ASCED broad categories) for 55 categories of post-school qualification. The 56th group is for no post school qualification.
The VUEF extension identifies 97 occupations, consistent with the ANZSCO minor group (3-digit) classification.
The key data requirements for the VUEF extension are the following matrices:
- skill by occupation (headcounts);
- skill by occupation (wage bill);
- industry by occupation (headcounts); and
- industry by occupation (wage bill).
The last of these is necessarily consistent with industry wage bills in the Vic-Uni model input-output database.
The data requirements for the estimation growth in the labour force by skill group are the Survey of Education and Work (ABS 6227.0), the census, and ABS population projections by age and sex.
To summarise, the Vic-Uni model with the VUEF extension is used to produce annual forecasts over an eight-year horizon of employment cross-classified by industry (115 input-output industries) and occupation (97 ANZSCO minor groups), and by skill (56 groups as described) and occupation.
These forecasts are processed through several auxiliary programs to:
● generate regional estimates;
● reclassify industry forecasts from the input-output industries to ANZSIC class;
● disaggregate ANZSCO 3-digit forecasts into ANZSCO 4-digit (Unit group) forecasts;
● estimate net replacement rates;
● classify employment into demographic groups; and
● classify employment by hours worked.