A nice write up. I agree that they are not making a comeback. As a pension actuary, I would love it. As someone with a reasonable knowledge of finance, it would be financial malpractice for a public company to restart a defined benefit plan.
I think many would agree that a plan sponsor can offer a similar level of retirement income for a cheaper cost than a participant driven defined contribution plan. However, it makes no sense for the plan sponsor to take on both mortality & investment risk. Beyond that, current accounting standards create too much income volatility for a defined benefit plan.
Thanks very much for the comments. Couple of things:
I think "financial malpractice" is a little too harsh. For example, it probably made financial sense for IBM's shareholders, at least temporarily.
Also, I assume in the second paragraph that what you mean by "cheaper cost" is that the employer will get better investment results (right?). I don't believe that -- it's possible to get passive investments at a very low cost in the 401(k), and I don't believe that most employers will beat passive, even by adding alts, for example.
Finally, I understand that accounting cost drives a lot of behavior, though I believe that economics should trump accounting when it comes to making business decisions. I once had the CFO of a client (a large auto supplier that was doing well while its peers were going bankrupt during the financial crisis) stop me in the middle of explaining how, even though something made sense, there were negative accounting implications; he said (as I remember it), "Never mind the accounting. Do the right thing and the accounting will take care of itself." I wanted to kiss him (but refrained).
Thanks again -- let me know if I misunderstood something or if you think I'm wrong about anything.
Larry - my comment about "cheaper cost" is simply that the plan sponsor can/should get better investment results than a 401k participant as they theoretically have professional management and should be less likely to swing with the emotions of the day.
While I agree that economics should trump accounting in the real world (and privately held companies are more likely to follow that), they need to respond to shareholders so managing EPS is important whether we like it or not. As actuaries, we may be able to understand that changes in discount rates do not impact the long-term cost of the plan, I'm not so sure that Wall Street and shareholders have the same view.
Great summary and covers items largely missed in the popular media. I would include an addendum that long-term investments in stocks is in many ways less risky that long-term investment in bonds and was one way that some employees were able to garner returns from classic DB plans and 401(k) plans. This is implicit in your discussion of allowing employees to bear risks according to their preferences and in the end discussion on tontines.
The way the law and the courts allowed conversion to cash balance but not 401(k) has in some ways stymied broader reform of private pension policy for decades. Yet I wonder if in fact we aren't mainly through that transition period and whether some new reform of pension policy couldn't be resurrected--one that tried at some level to include some of the best features of DB, DC, and cash balance.
Finally, a question. Does or could IBM also adopt an opt-out add-on 401(k) plan that would allow many employees to garner the higher long-term returns from stocks?
Thanks very much for the thoughtful comments. I'll start by trying to address the question, and then blather about other stuff.
I'm sure the IBM 401(k) is still open for employee contributions which employees can probably invest however they want (like existing 401(k) balances) from the plan investment menu. It's just the future employer contributions that got shifted to cash balance. I believe there's nothing legally stopping IBM from offering a choice between the two for future contributions, but if they did that, most employees would probably choose the 401(k), which would defeat the purpose of using up the DB surplus. I have to believe that IBM didn't do the shift primarily because they thought it was better for employees, but rather because it's better for them, and, they would probably say, doesn't hurt employees, at least not much.
I agree that employees should be allowed to invest in stocks, and probably should in most cases, and think that target date funds are an appropriate default to allow people to do so without needing to pay attention or know a lot about investing. But I also think if people are really risk averse, they should be allowed to invest in Treasuries or something else low-risk.
DB plans, including most cash balance plans (like IBM's), are more complicated, though. In traditional DB plans and most cash balance plans (including IBM), it makes no difference at all to the employee how assets are invested. The employee's benefit is the same and the employer is on the hook regardless of investment performance. With any given corporate DB plan, there's a good case that sponsors shouldn't be investing at all in stocks (though most do): http://users.erols.com/jeremygold/154hud.pdf (Jeremy Gold -- Gentlemen Prefer Bonds. The late Jeremy Gold was an actuary and PhD economist.)
I'd question that employers presumably getting higher returns from stocks means they can offer a higher benefit level in a DB plan, if that's what you were saying. DB pensions, as a very low-default-risk contractual promise, have the same value -- both as an asset of its employees and a liability of its shareholders -- regardless of how assets are invested. If the employer invests in stocks, it is creating an asset-liability mismatch. To say this another way, if the employer replaces $1,000 of cash (or other) compensation with $1,500 worth of pension promise because it thinks it can pay something worth $1,500 today by investing $1,000 in stocks, it is in fact increasing pay by $500 and costing its shareholders $500, and could, equivalently, just raise pay by $500 instead.
I agree that in the private sector, we're pretty much through the transition from traditional DB to DC, and not going back.
Regarding the best features of DB and DC and cash balance: Cash balance in some sense was an attempt to combine DB and DC, and isn't a good design for a number of reasons in my opinion. My preferred combination of DB/DC features is DC accumulations invested as people want, with a good default for people who aren't informed or don't pay attention, combined with the opportunity (not requirement) to pool longevity risk and receive lifetime income during retirement in a tontine invested in accordance with the participant's risk preference to provide some combination of lower stable income vs. varying income with some upside potential but downside risk.
I feel like current 401(k) DC accumulation is pretty good, and that the remaining step is to legally facilitate longevity pooling in tontines so people don't have to buy insured products, which is where current law steers them owing to the efforts of the insurance lobby.
This would be pretty close to Australia's approach conceptually, except that the accumulation vehicles -- superannuation funds -- are not employer-affiliated, I understand, but independent entities to which employers contribute (which Shlomo Benartzi advocates: https://www.wsj.com/personal-finance/retirement/update-401k-retirement-workers-f9abb373?st=U6rdzv&reflink=desktopwebshare_permalink). There are one or two "supers" that offer optional tontines for decumulation (not sure if there are alternative investment options -- I don't think so). This approach makes a lot of sense to me.
Please tell me if I'm missing something or am wrong about anything.
Thanks again for reading and taking the time to respond -- really appreciate it!
A nice write up. I agree that they are not making a comeback. As a pension actuary, I would love it. As someone with a reasonable knowledge of finance, it would be financial malpractice for a public company to restart a defined benefit plan.
I think many would agree that a plan sponsor can offer a similar level of retirement income for a cheaper cost than a participant driven defined contribution plan. However, it makes no sense for the plan sponsor to take on both mortality & investment risk. Beyond that, current accounting standards create too much income volatility for a defined benefit plan.
Thanks very much for the comments. Couple of things:
I think "financial malpractice" is a little too harsh. For example, it probably made financial sense for IBM's shareholders, at least temporarily.
Also, I assume in the second paragraph that what you mean by "cheaper cost" is that the employer will get better investment results (right?). I don't believe that -- it's possible to get passive investments at a very low cost in the 401(k), and I don't believe that most employers will beat passive, even by adding alts, for example.
Finally, I understand that accounting cost drives a lot of behavior, though I believe that economics should trump accounting when it comes to making business decisions. I once had the CFO of a client (a large auto supplier that was doing well while its peers were going bankrupt during the financial crisis) stop me in the middle of explaining how, even though something made sense, there were negative accounting implications; he said (as I remember it), "Never mind the accounting. Do the right thing and the accounting will take care of itself." I wanted to kiss him (but refrained).
Thanks again -- let me know if I misunderstood something or if you think I'm wrong about anything.
Larry - my comment about "cheaper cost" is simply that the plan sponsor can/should get better investment results than a 401k participant as they theoretically have professional management and should be less likely to swing with the emotions of the day.
While I agree that economics should trump accounting in the real world (and privately held companies are more likely to follow that), they need to respond to shareholders so managing EPS is important whether we like it or not. As actuaries, we may be able to understand that changes in discount rates do not impact the long-term cost of the plan, I'm not so sure that Wall Street and shareholders have the same view.
Understood. Great to have your perspective, Parker. Thanks again.
Great summary and covers items largely missed in the popular media. I would include an addendum that long-term investments in stocks is in many ways less risky that long-term investment in bonds and was one way that some employees were able to garner returns from classic DB plans and 401(k) plans. This is implicit in your discussion of allowing employees to bear risks according to their preferences and in the end discussion on tontines.
The way the law and the courts allowed conversion to cash balance but not 401(k) has in some ways stymied broader reform of private pension policy for decades. Yet I wonder if in fact we aren't mainly through that transition period and whether some new reform of pension policy couldn't be resurrected--one that tried at some level to include some of the best features of DB, DC, and cash balance.
Finally, a question. Does or could IBM also adopt an opt-out add-on 401(k) plan that would allow many employees to garner the higher long-term returns from stocks?
Thanks very much for the thoughtful comments. I'll start by trying to address the question, and then blather about other stuff.
I'm sure the IBM 401(k) is still open for employee contributions which employees can probably invest however they want (like existing 401(k) balances) from the plan investment menu. It's just the future employer contributions that got shifted to cash balance. I believe there's nothing legally stopping IBM from offering a choice between the two for future contributions, but if they did that, most employees would probably choose the 401(k), which would defeat the purpose of using up the DB surplus. I have to believe that IBM didn't do the shift primarily because they thought it was better for employees, but rather because it's better for them, and, they would probably say, doesn't hurt employees, at least not much.
I agree that employees should be allowed to invest in stocks, and probably should in most cases, and think that target date funds are an appropriate default to allow people to do so without needing to pay attention or know a lot about investing. But I also think if people are really risk averse, they should be allowed to invest in Treasuries or something else low-risk.
DB plans, including most cash balance plans (like IBM's), are more complicated, though. In traditional DB plans and most cash balance plans (including IBM), it makes no difference at all to the employee how assets are invested. The employee's benefit is the same and the employer is on the hook regardless of investment performance. With any given corporate DB plan, there's a good case that sponsors shouldn't be investing at all in stocks (though most do): http://users.erols.com/jeremygold/154hud.pdf (Jeremy Gold -- Gentlemen Prefer Bonds. The late Jeremy Gold was an actuary and PhD economist.)
I'd question that employers presumably getting higher returns from stocks means they can offer a higher benefit level in a DB plan, if that's what you were saying. DB pensions, as a very low-default-risk contractual promise, have the same value -- both as an asset of its employees and a liability of its shareholders -- regardless of how assets are invested. If the employer invests in stocks, it is creating an asset-liability mismatch. To say this another way, if the employer replaces $1,000 of cash (or other) compensation with $1,500 worth of pension promise because it thinks it can pay something worth $1,500 today by investing $1,000 in stocks, it is in fact increasing pay by $500 and costing its shareholders $500, and could, equivalently, just raise pay by $500 instead.
I agree that in the private sector, we're pretty much through the transition from traditional DB to DC, and not going back.
Regarding the best features of DB and DC and cash balance: Cash balance in some sense was an attempt to combine DB and DC, and isn't a good design for a number of reasons in my opinion. My preferred combination of DB/DC features is DC accumulations invested as people want, with a good default for people who aren't informed or don't pay attention, combined with the opportunity (not requirement) to pool longevity risk and receive lifetime income during retirement in a tontine invested in accordance with the participant's risk preference to provide some combination of lower stable income vs. varying income with some upside potential but downside risk.
I feel like current 401(k) DC accumulation is pretty good, and that the remaining step is to legally facilitate longevity pooling in tontines so people don't have to buy insured products, which is where current law steers them owing to the efforts of the insurance lobby.
This would be pretty close to Australia's approach conceptually, except that the accumulation vehicles -- superannuation funds -- are not employer-affiliated, I understand, but independent entities to which employers contribute (which Shlomo Benartzi advocates: https://www.wsj.com/personal-finance/retirement/update-401k-retirement-workers-f9abb373?st=U6rdzv&reflink=desktopwebshare_permalink). There are one or two "supers" that offer optional tontines for decumulation (not sure if there are alternative investment options -- I don't think so). This approach makes a lot of sense to me.
Please tell me if I'm missing something or am wrong about anything.
Thanks again for reading and taking the time to respond -- really appreciate it!