Geoff Gannon November 8, 2016

The Possibility of Negative Interest Rates in the U.S.

Someone who reads the blog emailed me this question:

“What’s your general thoughts on negative interest rates? Do you just take that out of your consideration for US banks?

I have read about some interesting analysis from smart people that outlines their reasoning for a much lower interest rates going forward than the past, say, 50 years. Most include points like less capital needed ahead because of the lower capital of new tech companies and more savings from around the world with people living longer.

Do you have any thoughts here?”

Yes. I do just take negative interest rates out of consideration when analyzing U.S. banks. I wrote reports on Prosperity (PB), Frost (CFR), Bank of Hawaii (BOH), Commerce (CBSH), and BOK Financial (BOKF). I can’t remember discussing negative interest rates in any of those reports. And they weren’t brief reports. I did discuss a lot of scenarios – like interest rates staying lower longer than I expected. But negative interest rates weren’t discussed.


I think talk of negative interest rates is the result of years and years of low interest rates. People talk a lot about the long-term average price of a barrel of oil, a house, etc. in the early stages of a bubble. Later, when prices have been out of whack for years and years they stop listening to the people who say you’re going to see mean reversion. Instead, they believe the now is normal. When the now is new – people know it’s abnormal. But once the now – in this case, rates near zero – goes on long enough, the recent past erodes their memory of the distant (like 70 years ago) past.


It’s natural for people to project the recent past into the future. Rates are low now. But they were low just after World War Two as well. Look at long-term corporate bond rates in 1946. They were low. Very short-term government bond yields (and the Fed Funds Rate) have been lower in the past few years than at just about any time in the past. But, I’m not sure the things that would matter more to an investor – things like the long-term corporate bond yield – are lower now than they were 70 years ago. And, of course, in between you had some very high yields. A long time ago, I wrote a series of posts about normalized P/E ratios. These are like the Shiller P/E ratio. I looked at normalization over periods longer than the 10 years he focused on. But, I don’t think the result is that different whether you are using a 10-year average, 15-year average, or 30-year average of past earnings. What I found is that the stock market – in normalized P/E terms – tended to get more expensive for about 17 years and then tended to get cheaper for about 17 years. For example, the market (I used the Dow) reached a peak normalized P/E ratio in 1965, it hit the bottom in terms of normalized P/E in 1982, and then it peaked again in 1999. If we went further back in time, the pattern wouldn’t be that different. You have 1929 (peak), 1942 (valley), 1965 (peak), 1982 (valley), 1999 (peak). Buying near the peaks and holding wasn’t such a good strategy. Buying near the valleys and holding was a very strategy. But, the problem is – of course – that we are talking about 13 years between 1929 and 1942 in which the market tended to have a contracting normalized P/E multiple. Then you had 23 years from 1942 to 1965 where the multiple tended to expand in normalized terms. Then 17 years of contraction. And then 17 years of expansion. So, we have “cycles” that are 13-23 years in terms of just one side of the cycle. It would take 26-46 years to have a complete cycle of extraordinarily low normalized P/E and extraordinarily high P/E. Many investors will only live through one or two years like 1929, 1965, and 1999. Likewise, many investors will also only live through one or two years as cheap as 1942 and 1982. A trend that continues for a decade or more is hard for people not to get caught up in. People got caught up in the dot com bubble and that was just the tail end of a rise in stock prices from 1982 to 1999. People got caught up in the housing boom – that was like half a decade. The recent oil boom lasted maybe 10 years. Interest rate cycles are at least as long. So, while I am sure there are good arguments for why interest rates might be negative for a long time – there were also arguments made for stocks in 1999, houses in 2006, and oil a few years ago.


I think there are plenty of arguments against negative interest rates being normal. For one thing, in the U.S. the rate of population growth is fine because of immigration. So, you have a very stable population pyramid. Now, that’s not true in some other countries. China is going to have a contracting pool of workers. I think it would make more sense if someone was arguing that China’s long-term future included very low interest rates than arguing that the U.S.’s future included very low interest rates. The U.S. has relatively high returns on assets. So, if a company gets paid nothing to put money in the bank – it should expand or if the incremental ROA is terrible because there’s nothing to do, it should buy back its stock. Eventually, the prices of stocks, buildings, etc. would be pushed too high and this wouldn’t be a way to use cash. But, it would take longer than it did for a similar situation to develop in Japan during that country’s growth phase or in China today. Companies in those countries have lower returns on assets. It is easier to imagine a situation where households and companies are leaving money idle if their alternative is a low return on asset investment. Countries like Japan and China have long histories of lower returns on assets than the U.S. You’d need the ROA in the U.S. to decline closer to wherever you expect interest rates to stay. Otherwise, it would be too obvious an action to borrow and do anything with the money.


You mention things like more savings around the world. I agree with that. In Asia, you have high percentages of workers relative to the total population. That leads to higher savings. So, the total amount of savings is going to depend in part on the ratio of workers to dependents. If a country like China, Taiwan, or South Korea has a lot of workers relative to its total population today – a dependency ratio of less than 1 to 1 – then it is going to have more savings now and less savings in the future. But this is a temporary – cyclical – phenomenon. It’s like borrowing money now to make investments. You can increase leverage today. But that will cause you to decrease leverage in the future. So, you are growing faster than your sustainable rate of growth now and then you will have to grow slower than your sustainable rate of growth in the future. Returning to the concept of ROA – if a company grows its business faster than its ROA, that company is pursuing an unsustainable rate of growth. That doesn’t mean it can’t grow in the future. But, it does mean it has to grow slower in the future relative to its ROA – because it can’t keep growing debt faster than it grows assets. I see that as a potential problem in places like China – because they have a lot of firms growing faster than their ROA. I don’t see that as a potential problem in the U.S. In the United States, companies fund themselves primarily from retained earnings. They do this even though their return on assets is often higher than what the cost of debt would be. For example, a company might be able to earn a 6% after-tax ROA (9% pre-tax) and yet borrow at 6% pre-tax (4% after-tax). Retained earnings are not the theoretically best way to fund the company. It would seem to make sense to borrow at 6% and benefit from the tax savings. U.S. companies don’t do this much. And I think it’s because that is not the analysis most companies do. They simply look at everything they want to do and if they can fund it out of retained earnings – they do. Some companies are different. John Malone is going to run a company he controls more rationally than that. He’s going to use as much debt as possible if using debt is cheaper than using retained earnings.


I don’t have any problem with the idea of very low long-term interest rates. But, I do have a problem with assuming they would happen even in situations where you had not already made some pretty severe mis-investments as an economy. I have trouble reconciling high returns on assets and low interest rates. So, if the firms in an economy are already earning low ROAs and then they start borrowing a lot and investing that borrowed money in low ROA activities – yes, I can see how you would end up in a situation where rates are close to zero for a very long time. I don’t see how that would happen unless you invest in low ROA activities though.


Now, interest rates – at least the short-term ones that matter a great deal to banks – are set by institutions like the Federal Reserve. And they may have goals like full employment that encourage them to set rates lower than would make the most sense in terms of driving good long-term investment decisions. For example, they might see a pool of unemployed workers. And they might know land prices are high. But, they know that if really low rates will stimulate homebuilding activity the stimulation such homebuilding will give the labor market right now is big. And so encouraging the building of a home that will turn out to cost more now than it will be worth in the future isn’t a concern for them. I’m not saying some people might not want negative interest rates at time certain cyclically difficult point in time. But, keeping them there long enough to change the investment thesis on a bank, insurer, etc. – is a taller order.


Let’s look at why. First, we have to consider that all banks, insurers, etc. have a P/E ratio. And that if interest rates are low that will tend to both decrease what they could earn on their investments – so their ‘E’ – but it will also tend to increase the P/E multiple that investors award all stocks. This includes the P/E multiple of banks, insurers, etc. Let’s say Progressive usually makes 5% on its investments. Interest rates fall so Progressive is now making 2.5% on its investments. But, if the normal earnings yield Progressive had traded at was 5% (a P/E of 20) and it now drops to 2.5% (a P/E of 40) you would have offset the difference in terms of the investor’s return in the stock. If investors expect less return from all their assets – that means banks, insurers, etc. make less on their loans, bonds, etc. But it also means stock investors are willing to make less. So, you can sell the stock for more. Now, Progressive would be hurt by this situation. But Progressive makes some of its money from underwriting and some of its money from investing. Banks also make income from fees, etc. If they aren’t making money lending out checking deposits – they can charge ATM fees, monthly fees, etc. on those accounts so they don’t lose money in a zero-interest environment. They can also buy long-term bonds.


Banks can also make longer-term commitments. This is the risk I am most worried about. And I did talk about it a little in the Bank of Hawaii report and a lot in the Frost report. These banks don’t just make loans. They invest about half their balance sheet in securities. When short-term rates are low, these banks tend to invest in longer and longer term securities. Instead of buying a 5-year bond they buy a 10-year bond. This exposes them to interest rate risk. If interest rates rise, the market value of the securities on their balance sheet will fall. They know this. They talk about this. And, honestly, for banks like BOH and CFR that benefit so much from higher short-term interest rates – it’s mostly just an initial offset. They will immediately write-off a lot of value. But then their cash earnings will start rising. It won’t threaten their solvency.


So, no. I’m not worried about negative interest rates. But, I am worried that short-term interest rates have been so low for so long that banks and insurers have been reaching further and further out in terms of the length of the commitments they are making. I’d rather they didn’t do this. I’d rather Frost kept more money at the Fed earning nothing and less money buying longer term bonds. Doing that would weigh down today’s EPS. And it probably wouldn’t turn out better for them in the long-run than buying overpriced bonds now. But, I think it’s a bad idea to ever buy an asset you know is overpriced. And these banks know the bonds they are buying are overpriced. They just don’t want to keep billions of dollars idle at the Fed for years.


Banks like BOH and Frost are interest rate sensitive. So, there’s a huge offset built into their business model that will mitigate the losses their securities portfolios will have. That’s not true for insurers. A lot of insurers have bond portfolios that are too expensive and too long-term. So, I’d suggest avoiding most insurers till interest rates are much higher again. Sure, I’m worried that it might take a lot longer for the Fed to raise rates than I’d expect. But I’m not worried that rates will be sustainably negative at any point in the U.S.