More on Conference Board Model of Corporate Tax Cuts
Further to my post yesterday about how the Ontario PCs have vastly overstated their own consultants’ estimates of the number of jobs produced by their various policy proposals (including lower corporate taxes, lower electricity prices, interprovincial free trade, and regulatory reduction), some have asked me about precisesly how the Conference Board report simulated the corporate tax reduction I was discussing.
At the bottom of p.8, their report (available on-line here) indicates they are simulating a reduction in corporate taxes of 1 percent of the corresponding tax base (ie. pre-tax corporate profits), which is equivalent to a one-point reduction in the corporate tax rate. Table 5 indicates that will reduce revenues (and enhance after-tax corporate cash flow) by $532 million in the first year, rising to $850 million after 10. That impact strikes me as somewhat small, but it is in the ballpark. (Total Ontario corporate rax revenue at the current rate is about $11 billion; that reflects the application of the current rate to pre-tax profits, less the various deductions companies are allowed.)
They then report the estimated changes in levels of GDP, employment, and other variables to arise from that tax reduction.  Table 5 indicates that employment is 5323 higher after 8 years. The Conference Board says the 1-percent impact can be multiplied linearly for tax cuts of different sizes (which is roughly true, to a point), so for the 3.5 point cut proposed by the Tories that number should be multiplied by 3 (to 18,631). For reasons I work through in detail in yesterday’s post, the Tories misinterpreted the 10-year cumulative person-year employment increment reported by the Conference Board at the right side of Table 5, calculated an average yearly figure for that, and then added that many new jobs to employment in each year of their job plan. That’s clearly wrong: having a job and keeping it for 8 years, doesn’t mean you have actually created 8 jobs. There’s no other way that they could have come up with the 14,976 annual job gain (for an 8-year total of 119,808) reported in their technical backgrounder, which I replicate precisely using my methodology.
It is apparent that the Conference Board was not simulating a “1 percent decline in corporate tax revenues.” They explicitly indicate (p.8) that they reduce taxes by 1 percent relative to the base (not 1 percent relative to the previous level of taxes). The $532 million first-year impact cited in Table 5 is far larger than 1% of current CIT revenues. Moreover, it would be unusual to simulate a policy change by imposing a certain increment in an endogenous variable (the amount of revenue raised by the CIT depends on economic activity, profits, and all the other variables determined within the model; it can’t be specified in advance). It would be possible to simulate a “1 percent reduction in the tax rate” (reducing it from 11.5% to 11.385%) but that would be unusual, and at any rate the impact effect reported in Table 5 is far larger than this.
It is clear that the Ontario PC backgrounder counted at least 100,000 too many jobs from their proposed CIT reduction, even on the basis of their own consultant’s report. Across other policy measures listed in their backgrounder (lower electricity prices, interprovincial free trade, regulatory reductions, PIT cuts), there are at least 100,000 more unjustified jobs contained in their million-job tally resulting from a similar error. I cannot replicate the Tories’ job estimate from PIT reductions (support for which they also cite the Conference Board report) in the same way I have explained where their CIT estimate came from; but no matter how one interprets the Conference Board’s simulation of PIT reductions (it is not stated in that case exactly what was reduced by 1%, unlike the CIT experiment), they have still clearly counted far more jobs on that issue than the Conference Board report suggests. It’s also immediately apparent that they claimed eight-years of person-year employment for the PIT reductions (the technical backgrounder claims 47,080 in total from 8 years of annual 5885 job gains), even though they only propose to reduce PITs in their dreamed-for second term (after the deficit has been eliminated). That’s another pretty big and obvious error.
I am still waiting for the PCs to publicly explain in detail how the job-creation numbers in their technical backgrounder were formulated. I think I am going to be waiting a long time. I cannot find the technical backgroundere on their web site anymore (again, we are all indebted to the Ottawa Citizen’s David Reevely, both for posting the documents and for being the first one to identify the multiple-counting problem); they would clearly prefer to simply change the subject (eg. to gas plants).
On second reading there are other interesting aspects to the Conference Board simulation of corporate tax reductions. The one that jumped out at me was their estimate of increased business capital spending after the tax cut (reported in Table 5, and the main driver of economic benefits in the simulation), reported in the fifth line of Table 4. They see an additional $133 million of business investment in the first year, rising to $227 million in the third year. In other words, by their estimates, less than one dollar in three of the CIT cut is reinvested by business in new fixed capital investments. This highlights the problem that has been experienced with CIT reductions as a stimulative tool. They translate only weakly into new business spending. That’s why the final gain in GDP (even counting indirect and induced multiplier effects) is always smaller than the initial cost of the tax cut. Even in the Conference Board study, one big lasting legacy of CIT cuts will be an additional increment to corporate cash hoarding, worth over $600 million per year by the 10th year (comparing the value of the CIT reduction in that year to the modest increase in capital spending). That sounds like a good reason not to do it at all.
Remember also that the Conference Board report did not incorporate (at the PCs’ request) the negative effects on GDP of employment from any offsetting reduction in other government programs (which the PCs have promised they would do, making the CIT cut supposedly “revenue neutral”). They make this clear on p.5. It is thus not a reasonable simulation of what the party is actually proposing.
Those problems, however, may pale in comparison to the big math error.