Kevin Milligan

Comments on Liberal Fiscal Plan and Costing
Professor Kevin Milligan, Vancouver School of Economics
University of British Columbia
September 26, 2015

In this blog post, I am providing an assessment of the Fiscal Plan and Costing document circulated today by the Liberal Party of Canada.

Disclosure and Motivation

As outlined in my disclosure, I have been an occasional unpaid advisor to the Liberals in the past few years and in that capacity have given advice on some of the individual items in the costing document. In addition, a few days ago I gave some feedback on a draft of this Fiscal Plan. As outlined in my disclosure, I am a member of no party. I am endorsing no platform.

Through this election cycle, I have also responded to requests for help with economic advice from the Conservative Party and the NDP. I view these engagements with policymakers as part of my role as a professor at a public university.

In the future, I think the economic discussion during election campaigns in Canada would be much improved if there were common forecasts and credible platform costing by all parties. I’m glad to see both the NDP and the Liberals proposing a change to the Parliamentary Budget Officer mandate to allow these services to be offered in future election cycles. In my view, this would be an improvement over the status quo.

But for now, many voters may struggle without some kind of expert assessment of party platforms. That’s why I decided to write this post—I’m doing this on my own accord; it was not requested of me. Having disclosed my involvement in the process that generated the Liberal document, I leave it up to the reader to decide if my writing is impartial.

Overall Assessment

The costing strategies used in the Liberal Fiscal Plan meet very high standards and the budget balance targets are achievable.

Below I discuss five aspects of the document in detail.

1. Forecasting framework

The core of the forecast is built from the projection provided by the Parliamentary Budget Officer for the years 2015-16 to 2017-18. This provides the ‘revised forecast’ in the first two columns of the ‘planning framework’ on p. 7. The last two columns show the forecasts for 2018-19 and 2019-20, with a small discount relative to the April budget.

Projecting GDP (and the resulting spending/revenue/budget balance) three and four years out involves a lot of uncertainty. While macro-forecasting is not my expertise, I don’t think the numbers used for the 3rd and 4th year are unreasonable.

Over the four years of the framework, the April budget forecast had a total of $18.7B in combined surplus and contingency. In the Liberal document, the four-year total is reduced to $13.8B, a reduction of $4.9B. Given the reduced growth estimates by the Bank of Canada and the July PBO report, I believe the Liberal planning framework uses a prudent forecast.

I’d like to make two important points about this forecast

First, the small deficits cause no structural fiscal concern. The deficits in years 1 and 2 of the plan are only about 0.5% of GDP. In each year of the plan, the debt-to-GDP ratio declines. This ratio is used by economists as the best gauge of long-run fiscal sustainability.

Second, the forecasting framework does not make use of ‘multipliers’ in the GDP or budget baseline numbers. Unless every dollar of the deficit crowds out other activity in the economy, the economy is very likely to grow in response to the small deficits—at least to some degree. This extra growth would generate extra tax revenue relative to the assumed amounts in the planning framework.

The multipliers estimated by the Department of Finance are shown on p. 7 of the Liberal document. If one took those multipliers and cut them by half or even two-thirds, the impact on the bottom line would still be counted in the billions of dollars.

The Liberal document has followed Department of Finance practices and not included these multiplier effects. In my view, this provides a fairly large fiscal ‘cushion’ that should be kept in mind when evaluating other parts of the document.

2. Overall costing strategy

A variety of strategies were used for costing. External validation from the Library of Parliament was used extensively. Standard tools of the trade like the SPSD/M model were also used. Other line items come from the CRA, Finance Tax Expenditures, and the Public Accounts. Finally, some of the bigger items were costed using purpose-built fiscal-demographic models.

This overall costing strategy meets very high standards. I believe the appropriate source is used for each of the items I checked. In particular, the use of the Library of Parliament to check on the costing of several of the tax line items provides very credible numbers. This is a standard for transparency for which all parties should aim in costing their election promises.

3. Canada Child Benefit

In the spring, I had a chance to dig deep into the fiscal/demographic model the Liberals developed for the Canada Child Benefit. I developed strong confidence in the model, in part from its benchmarking against SPSD/M, Public Accounts, and CRA numbers.

The Canada Child Benefit would cost $21,725M in 2016-17, paid for by cancelling four existing measures totalling $19,955M. The difference of $1.8B represents the net draw on resources.

The model completely accounts for the taxation of the Universal Child Care Benefit, contrary to some press reports from May. Also, there was some concern in the press about a different (lower) forecast for the existing NCBS/CCTB/UCCB. However, changes that affect the forecast for those existing benefits would also affect the proposed CCB. The Liberals have shown me numbers suggesting that the new lower forecast would actually decrease the net draw on resources.

Overall, I think the costing of the CCB is very credible and easily meets the standards I expect from public policy analysis.

4. Expenditure and Tax Expenditure review

I normally approach unspecified ‘expenditure savings’ line items in budget documents with a great deal of caution. My caution is somewhat relieved in this Liberal document by two factors.

First, the Liberal document specifies more than a page-full of specific ideas on which the proposed review would focus. On the tax side, these include substantial items like the family taxation aspects of CCPCs, the taxation of options for high earners, and the tax expenditures for high earners specified by a PBO report. On the expenditure side, advertising and consulting are good targets for review, and count in the hundreds of millions of dollars.

Second, the timing of the revenue change is ‘backloaded’, not hitting $3B until the fourth year. This would give sufficient time for a thorough review to do its work and change administrative procedures to implement any changes.

I think that a serious expenditure and tax expenditure review should be able to reach the $3B target by year 4.

5. High Income Tax Bracket

It has been standard practice in Canada to account for potential rate changes by applying the new rates to the amount of income reported under the old rates. This kind of ‘static’ analysis does not account for the possibility that there may be some ‘behavioural response’ to the new tax rates. In the case of a high-income tax bracket, this behavioural response may take the form of tax planning by high-income tax payers who use make use of legal tax strategies to lower their tax bill.

The static estimate for the extra revenue from the new tax bracket for 2016-17 based on Library of Parliament numbers is $3.4B. The Liberal document only books $2.8B of revenue, allowing for up to $600M of behavioural response.

Accounting for behavioural response is always difficult because we can’t be sure of the magnitude. Economists (including myself) have tried to estimate these magnitudes, finding some evidence of sizeable response. Some economists might expect up to half of the potential revenue to disappear, but there are three mitigating circumstances here in the plan presented by the Liberals.

• This is federal taxation. Many estimates of high-income behavioural response are based on provincial taxation. There is a substantial difference in the scope for shifting income across provincial borders in response to a provincial tax change. This means that the response to a federal tax change will be much less than to a provincial tax change.
• The tax expenditure review. The review laid out by the Liberal document has a focus on the top 1% of earners. By shutting down potential tax planning avenues, the magnitude of the behavioural response will be diminished.
• Extra CRA resources. The Liberal document earmarks extra resources to the CRA for enforcement.

Taking this together I think that with strong administrative efforts and an effective tax expenditure review, 2.8B in additional revenue is a reasonable estimate.