Someone emailed me this question:
“I own Bank of America (BAC), Citigroup (C) and Wells Fargo (WFC, Financial). Each of them is growing deposits at different rates (off the top of my head, around 5%, 1% and 6% per annum). U.S. deposits have been very consistent at 6% for many, many years. This is probably an incredibly stupid question, but how can deposits continuously grow at such a clip? My initial reaction was that the Fed is printing money at a rate above real GDP growth of around 3.5-4.0%? Are deposits continuously pouring in from overseas? Is that growth rate sustainable / does it make sense (if it does, banks are obviously very attractive investments)?”
Growth in bank deposits is something of a special case because we are just talking about money here. It’s a very, very generic thing. So, we don’t have to worry about the popularity of money, possible substitutes, societal change, etc. This makes growth in banking far easier to predict than growth in any real good.
If a bank maintains the same market share it should be able to grow deposits at the rate of nominal growth in the local economy. This statement assumes that deposits as a percent of nominal GDP will neither rise nor fall. That’s not true cyclically. And it may not always be true even historically over the development of a state, country, etc. For example, American households don’t keep much of their savings in deposits because they have access to other types of investments. In many countries in Asia, that’s not true. People actually use deposits as a significant part of household savings. That changes the economics of banking. For example, it makes interest rates more important in attracting deposits from households. A lot of money kept in banks in the U.S. is not really surplus. It is money that is used for general operating purposes by either a household or a business that does not consider the account to be where they would actually put money they intend to save for many years to come. Having said that, there is still a cyclical aspect to deposits. You can see this historically. But, it may not be important to your analysis. If you are just looking for what the long-term rate of growth in deposits is likely to be – that’s easier to answer.
Let’s start with a theoretical explanation of what the growth in deposits – ignoring any cyclicality – should be in the U.S. banking industry overall. Banks are often local or regional. So, for most banks, the nationwide growth in deposits is irrelevant. As an example, I wrote reports on both Bank of Hawaii (BOH, Financial) and Frost (CFR, Financial). In those reports, I suggested that deposit growth could be as low as 3% in Hawaii and as high as 6% in Texas. That’s mainly because Texas should grow its population faster than the overall U.S. while Hawaii should grow its population slower than the U.S. This is largely just a matter of population density. Land is expensive in Hawaii. Land is cheap in Texas. People move inside the U.S. Not many people will choose to move from other U.S. states to Hawaii. However, many people will choose to move from another U.S. state to Texas. This explains almost all of the difference in expected deposit growth rates for Hawaiian and Texan banks.
So, what’s the theoretical description of what deposit growth should look like in a bank’s market area? Basically, it should be population growth plus real output per person growth (which is the real GDP number you mentioned) plus inflation. You said “My initial reaction was that the Fed is printing money at a rate above real GDP growth of around 3.5-4.0%”. Yes, nominal GDP growth tends to be higher than real GDP growth. The Fed talks about something like a 2% inflation target. So, if you expect 3% to 4% real growth, then you’d expect 5% to 6% nominal growth – if you believe that 2% target. Historically, the Fed – and almost every other central bank – has not tended to get inflation as low as its target. The long-term reality is more like 3% inflation despite talk of a 2% target. Recently, inflation has been quite low. Should you look at the more recent figure instead of the long-term figure? I have no way to choose between 2%, 3%, or 4% inflation as being more likely in the very long-term. You can use any of those numbers as reasonable. Let’s say the historical rate had been 3% and the target is 2%. Then why not use 2.5%? That’s as good a guess as any. No one knows what the long-term inflation rate will be. It is much easier to predict population growth. It’s also easier to predict real output per person growth. So, it’s easier to predict the long-term growth in real GDP than in nominal GDP. Let’s say U.S. population growth will be roughly 1%, real output per person growth will be roughly 1%, and then inflation will be roughly 2.5%. That gives you an estimate for nominal GDP growth of 4.5%. But, that number will be different in different states. Even if the numbers I just laid out hold true for the U.S. as a whole for many, many years to come – I’d still expect deposit growth in Hawaii to be less than 4% and deposit growth in Texas to be more than 5%. The difference in their rates of population growth should be at least 1%. As a rule, you’d expect places with low population density and a low cost of living to grow faster than places with high population density and a high cost of living if people were willing and able to move between the two places.
For bank deposits, it’s important to consider a few factors. One, different countries have different sized financial systems relative to their GDP. This is also true for states. The U.S. does not – for a rich country – have a particularly large amount of bank deposits and bank loans. So, it’s not something to spend a lot of time thinking about. You may want to spend more time thinking about cyclicality. The year-to-year rate of growth in bank deposits is pretty lumpy. It’s lumpy because you are looking at such short periods of economic growth and sometimes contraction that we are really dealing with something closer to acceleration and deceleration in the short-term rather than the long-term average “speed” of the economy. To smooth out this problem, don’t look at this year’s growth in deposits. Take something like a 7-year period. If you have data on a bank going back 17 years, then you have 10 years where you can calculate a 7 year period that ends in that year. There’s nothing magical about a 7-year period. You can use 5, 10, or 15 year periods. The difference between these periods will rarely be noticeable. But, sometimes, you have unusual economic events that will make some 5 year periods look strange.
If deposits are coming in at a cyclically unusual rate, you will probably notice a pattern inside the bank. You’ll see loans divided by deposits falling. So, you may see that a bank had $10 billion of deposits in 2006 and $9 billion of loans in that same year – but then in 2009, it had $12 billion of deposits and still only had $9 billion of loans. That means loans fell from 0.9 times deposits to 0.75 times deposits. This is typical of unusually fast deposit growth. You can also check this over the long-term by seeing if the company tends – over a couple decades – to have a consistent relationship between loans and deposits. Sometimes you’ll find a company had 60% to 80% of its deposits loaned out through the 1990s and early 2000s, but now has only 40% to 60% of its deposits loaned out in the 2010s. This is a sign that deposits are cyclically high and loans are cyclically low. Deposit growth has been higher than loan growth. At some point, loan growth will be higher than deposit growth. So, at some point, deposits are likely to grow slower than the nominal GDP of the region the bank operates in.
The growth of the industry’s deposits isn’t very useful to you. Three things matter more. One, the growth in deposits at the bank you’re looking at. Two, the growth in deposits per branch at the bank you’re looking at. And, most importantly, the growth in deposits per share of stock at the bank you are interested in. Companywide growth in deposits is close to irrelevant. At small sizes, there are economies of scale in banking. However, the greater efficiency of giant banks over big banks is almost entirely due to increased cross-selling and more fee generating businesses. There’s not much evidence that a bank with $200 billion in assets is more efficient in either gathering deposits or making loans than a bank with $20 billion in assets. This is not true at the branch level. A bank that has $150 million in deposits per branch will be much more efficient than a bank that has $15 million in deposits per branch. So, it’s always good to have deposits grow quickly and branches grow slowly. In fact, that is one really hopeful sign for banks – especially the big banks you mentioned in your question – for the future. They may be able to grow deposits at nominal GDP type rates while not growing branches at all anymore. Branches are being used less and less every year. This reduces the need for a bank to pay rent and salary relative to the amount of deposits (and therefore loans) it has.
I mentioned how I expect Hawaii to have very, very slow deposit growth over the long-term. That’s not as bad for Hawaiian banks as it sounds. I don’t expect any new entrants into the Hawaiian market. Mainland banks don’t really operate in Hawaii. I also don’t expect the four biggest Hawaiian banks (who together control virtually all deposits in the state) to open more branches than they close over the next decade or two. So, I expect the four big banks in Hawaii to have stable market share. That means even if I expect Hawaiian deposits to grow as slowly as 3% a year, I expect the nominal growth rate in deposits per branch to be at least 3% a year – and maybe higher. Furthermore, companies like Bank of Hawaii should earn a high return on equity while having little opportunity to re-invest in Hawaii. If they want to stay focused on that state, their only choice is to pay dividends and buy back stock. I expect BOH to buy back a fair amount of stock. Let’s say BOH buys back at least 2% of its shares each year. That means the growth in deposits per share will be higher than 5%. It also means the branch level economics should be stable to improving due to growth in deposits equal to or greater than growth in rent and salaries. Basically, it means the stock can return 5% a year or better plus whatever its dividend yield is. The dividend yield now is 2.4%. So, that’s a 7.4% total return potential on something as bad as 3% deposit growth. In the long run, a lot of “growth” stocks won’t manage to return 7% to 8% a year even while they grow their sales faster than the economy.
I should mention that fast growing banks are even better investments. Most of what I just laid out for Bank of Hawaii is equally true of a bank of equal quality that grows a couple percentage points faster in terms of its deposits. So, if you could grow deposits by 5% to 7% a year instead of more like 3% to 5% a year, the stock could return about 2% more a year. Again, banking is unusual in this case because money is a commodity type product but banks have no trouble retaining customers who deposit money with them. The only thing I can think of that’s similar is auto insurance. And auto insurance is much, much more price competitive than banking. In the U.S., banks don’t adjust the interest rates they offer with the same competitive aggressiveness that auto insurers adjust their premiums. It’s possible for a bank to retain about 90% of customers from year to year. In a normal year, any U.S. bank of real size that is run at all efficiently will have more cash profits than it needs to grow. That’s what makes U.S. banks – provided they are careful lenders – above average investment candidates. A lot of companies can grow 4% to 6% a year. However, there are banks that can grow 4% to 6% a year while paying out dividends and share buybacks of close to 4% to 6% of their share price at the same time. A business that grows 6% a year and pays no dividend isn’t attractive. A business that grows 6% a year and pays a 4% dividend yield at the same time is a market beating investment. For cyclical reasons, the banks you mentioned will grow their deposits somewhat slower in normal years than they have recently. However, they will be capable of having pretty high dividend yields in the future. The combination of growth in deposits per share and the dividend yield is what matters. For the U.S. as a whole, nominal GDP growth of 4% to 5% is possible. So, for big banks, deposit growth of 4% to 5% is possible. The question then is how much money they have – after retaining what they need to maintain leverage levels from year-to-year – to pay out to you. I don’t expect the S&P 500 to return 8% a year over the next 15 years. It’ll do worse than that. So, if you can find a bank that can – at a normal Fed Funds Rate – grow deposits per share by 4% a year while paying you a 4% dividend yield, you’ll beat the market.
I don’t own Bank of America (BAC), Citigroup (C), or Wells Fargo (WFC, Financial). However, Warren Buffett (Trades, Portfolio) does own Bank of America and Wells Fargo. And I certainly believe that Wells Fargo is capable of doing what I just described. I don’t own Wells because it’s more complicated than other banks I could find. But, I have nothing against the three banks you mentioned. I wouldn’t be surprised if they outperformed the market going forward. I especially think Wells Fargo is an excellent bank. In fact, I think Wells is the best bank I never wrote a newsletter issue about. Like I said, I didn’t reject it because it was a bad bank. I rejected it because it was more complicated than other banks. The only bank stock I own is Frost (CFR, Financial). It’s a Texas bank. I think it’s simpler than Wells. And I think I’d rather invest in Texas alone than the U.S. generally. But, I did newsletter issues on: Frost, Prosperity (PB), Commerce (CBSH), Bank of Hawaii (BOH, Financial), and BOK Financial (BOKF). A lot of them have much higher prices than when I picked them. But, I think they’re fine as a basket. I also see nothing wrong with the biggest banks like BofA, Wells, and J.P. Morgan (JPM) except that they’re complicated and they are subject to some special rules that regional banks (due to their smaller size) aren’t. But, honestly, I think just about any of the banks mentioned in this article are good. I think a basket of all of them is a perfectly good alternative to owning the S&P 500. The economics of a U.S. bank are better than the economics of the average publicly traded company. Banking in the U.S. is a good business. One caveat: nothing I’ve said here should be expanded into other countries. The U.S. banking industry is unlike banking in other countries. I’ve never found foreign banks I liked as much as U.S. banks – and I’ve definitely tried looking. In fact, I’ve wasted a lot of time trying to find foreign banks that were as good as U.S. banks. So, my advice is to stick to U.S. banks. And then try to figure out the long-term deposit growth per share potential plus the dividend yield. I think nominal GDP growth of the state (for a regional bank) or the country (for a nationwide bank) is a good proxy for long-term deposit growth. In the U.S., that means you’re probably looking at long-term deposit growth potential of maybe a little less than 3% to maybe a little more than 6% a year depending on which city or state the bank is in.
Disclosure: Long CFR
