Since my original post on the PBO Report (pdf), I have had a very useful email exchange with Chris Matier, the author of the Report. (It is to the PBO's credit that they really do want people to understand what they are saying, and are prepared to spend the time to help us.) This post is an update on my thinking.
Now I understand the PBO Report better, I think that people are being far too complacent about demographics and benefits to the elderly. The common reaction to the Report seemed to be "OK, so the PBO says it's all sustainable, so we can keep on doing what we are doing now, and there's nothing to worry about." I think that is the wrong conclusion to draw. I will explain why below.
My initial back of the envelope estimate was that, since demographics would nearly double the ratio of elderly (65+) to working age (15 to 64) by 2036, the total benefits paid by the government would also nearly double as a ratio to GDP. The PBO Report gave a much lower estimate. Chris Matier has re-run the numbers with different assumptions, and I now know where the differences are coming from.
It's all about the indexation of benefits. I don't mean indexation to inflation. We were both assuming full indexation to inflation. We were making different assumptions about indexation to GDP.
The Report assumes that benefits will be 50% indexed to GDP per capita. I was assuming that benefits will be 100% indexed to GDP per worker. That's two differences.
The 50% vs 100% difference is fairly simple to understand. If productivity is growing over time, and benefits grow at only half the rate of productivity, that would make benefits fall as a percentage of GDP, if demographics stayed the same. So that's one reason why the PBO Report gave a lower estimate for the total benefits/GDP ratio than my back of the envelope calculation.
The GDP per capita vs GDP per worker difference is a little more complicated. With an aging population, the number of working-age people will fall as a ratio of the total population, so GDP per capita will fall relative to GDP per worker. So if benefits are indexed to GDP per capita, rather than to GDP per worker, even if the indexation were 100%, that would make benefits a smaller percentage of GDP. So that's the second reason why the PBO Report gave a lower estimate for the total benefits/GDP ratio than my back of the envelope calculation.
You could ask: which assumption about indexation is more realistic, or more plausible? 100% indexation to GDP per worker (my assumption); 50% indexation to GDP per capita (the PBO assumption); or even no indexation to GDP at all, just indexation to inflation (the PBO's alternative assumption)? But I don't think that's the right question to ask in this case. I think it's better to ask: what assumption do people have at the back of their minds when they ask whether "what we are doing now" is sustainable with an aging population. What exactly do people mean when they talk about "what we are doing now"?
It's more a normative than a positive question. If you think that benefits to the elderly should maintain some given ratio between people's income when working and their income when they retire, then you are implicitly making my assumption. You believe that the OAS benefit should be x% of the average wage, where x is some constant that does not change over time. You believe in a fixed x% "replacement rate". You believe that benefits per old person should be indexed 100% to GDP per worker. And in that case "what we are doing now" is only just sustainable. Chris Matier estimates the "fiscal gap" drops from -0.9% of GDP to -0.2% of GDP when we switch from the Report's assumption (50% indexation to GDP per capita) to my assumption (100% indexation to GDP per worker).
0.2% of GDP is a little too close to 0.0% for my comfort level. If tax revenues fell by more than 0.2% of GDP, or if other expenditures increased by more than 0.2% of GDP, the whole federal budget would switch from being "sustainable" to "unsustainable". Not to mention the "unsustainable" provincial budgets in the face of increasing medical expenses for the old.
Maybe instead you believe that OAS benefits are too high now as a ratio of income per worker. And you want to see OAS benefits declining over time relative to income per worker, and you would prefer that to increasing the eligibility age from 65 to 67. OK. Then the PBO assumption fits with your normative views. But if you didn't believe that OAS benefits are too high relative to income per worker, and you don't want to see OAS benefits falling behind income per worker, you might want to reconsider the options.
I hope lots of people read this, Nick, it's such a clear explanation of the issues involved.
You wrote: "If you think that benefits to the elderly should maintain some given ratio between people's income when working and their income when they retire, then you are implicitly making my assumption."
Generally speaking, the literature on retirement talks about "replacement rates", i.e. an adequate pension income is one that replaces, say, 50% or 75% of pre-retirement income. So I'd say most often people thinking about retirement are making your assumption.
Posted by: Frances Woolley | February 17, 2012 at 09:43 PM
Thanks Frances! I was worried that I hadn't explained it clearly enough. (I just now edited it to stick in a sentence about the "replacement rate".)
Posted by: Nick Rowe | February 17, 2012 at 09:54 PM
In my random reading on OAS, I came on a good survey paper by Ken Battle. Ken had a concept "social policy by stealth" (something like that). He was talking about slowly abolishing things by not indexing them to inflation. I felt this was a bit like that, even if it's not exactly the same.
Now, you could make a case that OAS is too generous. And that it is better to cut OAS than to delay it till 67. And you might say it's better to cut it slowly, or just not increase it as productivity increases, so as to make the transition easier for people who have already made their plans. OK. But it does look a bit stealthy, if that's not what people had in mind.
Posted by: Nick Rowe | February 17, 2012 at 10:04 PM
Thanks Nick, that's really interesting.
Canadian spending on the elderly looks very low compared to the UK.
E.g. In the UK spending on old age pensions has increased from 4.6% of GDP in 2007, to 5.7% of GDP in 2012.
This definition excludes a number of benefits given to the elderly.
In Canada are there lobbying organisations for increasing welfare payments to the elderly?
Are you concerned that as the median voter ages, politicians will find it increasingly difficult to resist bribing retirees with increases in unfunded state retirement benefits?
For example, in the last election in the UK, a big part of the Lid Dems, (normally a minor party) election pitch was to increase the rate of uprating of the universal basic state pension. Politically, it was a pretty successful policy.
Similarly, in the recent election in Spain, despite all their problems, the conservative victor, Rajoy, promised to reinstate the indexing of state pensions. I believe state pensions consume an even larger share of Spain's government revenues than in the UK.
Posted by: Brit | February 17, 2012 at 11:21 PM
Brit: I'm really not sure how the Canadian pension system compares to other countries, like the UK.
This Report is just about Old Age Security, which is for everyone 65 and over (it's taxable income).
there's also Guaranteed Income Supplement (an income tested form of welfare for old people)
And the Canada Pension Plan, which you have to pay into while working.
"In Canada are there lobbying organisations for increasing welfare payments to the elderly?
Are you concerned that as the median voter ages, politicians will find it increasingly difficult to resist bribing retirees with increases in unfunded state retirement benefits?"
Yes and yes.
Posted by: Nick Rowe | February 17, 2012 at 11:29 PM
Two points. First, I think most people - who aren't economists - have been going along on the assumption that the OAS will mostly be kept up with inflation. I think that's what most people understand as what we're doing now.
Second, when you start talking about GDP per worker, I think you do have to consider that I think we expect that as the labour force shrinks in relative terms, we expect GDP per worker to rise, not just relative to GDP per capita, but relative to what the norm for GDP per worker is now.
Posted by: Jim Sentance | February 18, 2012 at 11:02 AM
Is that a reasonable expectation? It seems like a falling labour participation rate ought to drive higher productivity growth.
Posted by: Andrew F | February 19, 2012 at 02:19 AM
You're still chasing phantoms. There is no *automatic* increase linked to GDP, either per capita or per worker. OAS only goes up in real terms if legislation is passed to raise it.
You're basically arguing that we should cut OAS because...future politicians might raise it. Illogical.
And not even borne out by past actions. Take a look at http://www.servicecanada.gc.ca/eng/isp/statistics/cppstatbook/statbook2011.shtml#t4. Load it into a spreadsheet. Divide the benefits by the number of recipients and you have the annual benefit since 1980. Calculate each year's percentage increase. The numbers almost perfectly track inflation in Canada.
In other words, politicians haven't given OAS a real increase since at least 1980. Why would they start now, in the face of the baby boom?
No OAS crisis. Nothing to see here. These are not the droids you're looking for.
Posted by: tyronen | February 19, 2012 at 06:30 PM
Tyronen: "You're basically arguing that we should cut OAS because...future politicians might raise it. Illogical. ... why would they start now, in the face of the baby boom"
It's not as illogical as you might suggest. Think about the demographics. In 2030 1 in 3 eligible voters will be over 65 (and, once you factor in actual voting patterns, it'll probably be darn close to 50% of actual voters). That's a pretty hefty pressure group to push for increasing OAS/GIS. Why would politicians start increasing OAS in the face of the baby boom, the question answers itself, because the baby boom will be the most powerful interest group in Canada.
Push back the retirement age and you start cutting into the size of that group. Push back actual benefits, then at least the starting point for future increases will be lower.
And note, even if OAS were merely indexed to inflation, the PBO still estimates that Canada's finances as a whole are still unsustainable as a result of rising health care costs.
Posted by: Bob Smtih | February 21, 2012 at 09:25 AM
I will just have to keep reading on this subject because this does not clear up the argument for me. I am stuck on the relevance of GDP per worker vs GDP per capita. Why does it matter? Yes there are two ratios, but the GDP figure remains the same between the two ratios.
I tend to agree with the PBO that GDP growth will benefit from improved labour productivity stemming from a) forced innovation in Canada as labour becomes relatively scarce and b) importing of labour-intensive goods.
Posted by: Kathleen | February 21, 2012 at 09:58 AM
GDP per capita will tend to increase if GDP per worker rises, but that will be offset in whole or in part by the number of workers per capita falling. So GDP per capita will not grow at the rate GDP per worker does. If you tie the indexation of benefits to GDP per worker, the cost is going to go up much faster. Per capita it might not go up much at all. Significant implications I think for affordability.
In any case, I still think the practise and expectation of most is indexation for inflation.
Posted by: Jim Sentance | February 21, 2012 at 11:04 AM