Inflation, investment, and competitiveness.

I want to present an argument about business investment and inflation. To get to that though we need to take a step back and put in some context for what I want to argue. The basic argument for how firms behave is that they alter their variable costs so as to choose a level of production that, for a given price level, maximizes their profits. They’ll use those profits to, hopefully, pay the cost of capital on any long-run costs and then distribute the remainder to the owners of the firm. Obviously, it’s possible to get into a position where the price is such that no alteration of your variable costs will turn a profit and you’ll have to either find a way of reducing your long run costs or go bust. This is all graphed in a very traditional economics way by putting price/unit on the Y-axis and number of units produced on the X-axis. Then you put a nice straight line across indicating the market price and then you can plot your costs per unit to get both supply and demand. All very good, perfectly reasonable for considering short-run decisions/declaring bankruptcy/etc. You can put all the firms together and using their collective variable costs you can get a market price for the goods by comparing that with a demand slope for consumers indicating how many units of the good or service will be bought at a given price. Get the intersection of the two and you have a market price. Great.

What I want to talk about is the other half of that thought process- when is it sensible to expand and take on new fixed costs? Obviously, this isn’t an unexplored question in economics, I just think what I’ve put below is an interesting way of thinking about it that feeds into inflation and investment.

My argument is this: say you know the current market price per unit and you’ve already optimized your variable costs. You then want to decide whether to expand or not expand. You could then graph total costs including both the fixed costs you’ll incur by expanding and approximate variable costs against total income at the current price. For a given price, total income will be a diagonally sloping line, starting at zero for zero units sold and then sloping upward by the unit price. If you plot total costs against that then you can see, for any given expansion, whether it would be more or less profitable than now, at the current price, to make that investment. I’m assuming here that investments are priced as the interest paid on a loan of that amount. That’s true whether you do actually borrow the money or whether you use money you already have sitting around that could have been lent out at that level of interest.

I’ve created an arbitrary example and graphed it below. The optimum position on that graph in terms of profitability is 20 units. You can see from the bumps in the yellow line (total costs) that this is at the high end of capacity for two lots of investment. I.e if you were a firm that already had two machines/factories/whatever this investment represents you’d stick with those two. If you were at one you’d expand to two. If you were entering the marketplace as a new firm then ideally you would borrow enough capital to start with two machines/factories/whatever the capital investment represents here. Also, I know it’s not perfectly clear to see but I promise you every point on the yellow line is higher than any point to its left on the graph. You can see from the red line any spot where total profits after cost of capital go above zero. Just trust me that 20 is the highest peak.

So why go to the effort of graphing the situation differently? Especially when it creates a clearly silly-looking bouncy graph that doesn’t feel serious and "economicsy" at all?

Well, that bouncy graph makes a point that I think is quite relevant. I was listening to an Odd Lots podcast recently on Keynes and how one of the big points he makes in General theory is about how important stability of future expectations is to investment and how damaging uncertainty can be. This graph, after all, makes a key assumption: prices will be, at least, the current price.

Let’s go wild though and guess that future prices might be different from current ones. What does that do? Well, the income from zero units sold is always zero, that won’t change. What will change is the total income for each subsequent unit sold. If the price drops by £1 then, unsurprisingly, the total income from selling one unit will be £1 lower than the graph above and the income from selling fifty units will be £50 lower. I.e the right-hand side of the blue line will drop faster than the left-hand side. That makes a big difference to which allocation of capital is most profitable. Rather than the situation described above, with the current price, it’s perfectly possible to be in a situation with a new price where the rational place to be is just one machine/factory/whatever the capital allocation represents. A market entrant would only buy one, the firm already at one would stick, and the firm with two would struggle. Equally the reverse is possible, if prices rise it might be rational to have increased capital allocation beyond two machines/factories/whatever. That probably doesn’t sound like the most shocking conclusion in the world, it’s what you’d expect, but my point here is that that Keynes’ point around uncertainty is really important for investment and productivity.

In particular, what I’m saying is that if all firms could be confident that the balance of probability was that prices would rise in nominal terms then each would invest. However, if you lived in a world where 1. Price reductions, i.e deflation or low aggregate inflation, were common, and 2. You could not subsequently reduce your costs to capital, then there would be an obvious worry. If you work out that increasing your capital allocation is rational at the current price, and you also know that everyone else knows that, then you should expect everyone to increase their capital allocation, production to increase and because consumer preferences, presumably, have not substantially changed but there is more supply, prices will drop. If everyone else invests but you do not then prices will still drop, although not by as much, but they will have paid for new capital (and you won’t have) but you’ll be at the correct capital allocation. I.e. they pay, you gain! Presumably, everyone would know this as well and so rationally no one invests.

So let's talk about capital cost for a second. We can talk about a gold-based economy and a modern inflation mandated central bank. If you're in a world before fiat currency then you can imagine how this might exacerbate booms and busts. You're going to get a pretty wild ride as the cost of capital bounces around and prices are limited by gold supply/how quickly it can circulate. Animal spirits and expectations are going to run pretty rampant. Modern central banking works to, in theory, reduce this problem in two ways. Firstly if it achieves its inflation target then not every product will go up nominally in price but it at least mitigates the risk of substantial and sudden drops in prices. That makes investment safer. Also during deflation or low inflation central banks reduce the cost of capital. That can be helpful in mitigating the risk of having to put out interest payments on capital that is too expensive to afford in current economic conditions. Together those two things make investing less risky. It also perhaps helps explain why, despite textbooks saying lower interest rates spur greater capital investment by firms, changes in interest rates do not seem to change investment from firms, instead interest rate policy mostly works through personal borrowing and mortgages (house prices really do get pushed around by interest rates in a way firm's investment does not). This makes sense if you assume the reason the interest rates have dropped is an event that would have otherwise slowed investment or risen because of an event that would otherwise have increased investment. Investment isn't moving around precisely because central bank policy is mitigating the swings that would otherwise have happened. That's a good thing.

However, it isn't working perfectly and it's not clear that it works quite as uniformly as I describe there. It is true that when interest rates drop firms often refinance what they've borrowed at a lower interest rate. That reduces their fixed capital costs. However, it’s not true that this affects all firms equally. Just because the central bank lowers interest rates it doesn’t guarantee every firm can refinance their capital costs, nor does it mean that those who can refinance all get the same reduction in their capital costs.

Lenders have to be concerned not only about the generic cost of capital but the risk of any individual project. Borrowing costs are generally assumed to include both. So if a central bank were to reduce the generic cost of capital from 5% to 1% that doesn't mean a firm would go from paying 5% to 1% that's just the portion before risk. If risk adds 5% to both then it might be able to go from 10% to 6% borrowing costs. Another firm that only faces a 2% addition for risk would go from 7% to 3%. So the first firm has seen its costs reduce by slightly less than half, the less risky firm has had its costs reduced by slightly more than half. Before the second firm was paying 70% of what the first firm was paying, now it's paying 50%. Their risk hasn't changed, all that's happened is that the safer firm is at more of a relative advantage in financing.

I think this is a bit of a problem for monetary policy as it stands, as it means that not only did we start at high interest rates noticeably above inflation, effectively giving wealthy people free money (and more free money the more free money they had) but when we dropped interest rates that didn't really help as that pushed up the value of capital assets, which are mostly owned by those who already had wealth, and so that was free money for the wealthy too. Then we hit the zero lower bound on interest rates and started Quantitative Easing, which pushed up asset prices, giving free money to those who were already wealthy. When you add this to the point made above lower risk firms got an increasing advantage over less secure ones. In practice that means large, established firms got an advantage. Putting those two together is it remotely surprising that we now live in a world of increasing wealth disparity and a few big market-leading firms buying up or crushing smaller ones? Plus if inflation is now very low indeed below the 2% target, and the cost of capital might be low for big established firms but can't really drop much for small firms, then surely this helps explain part of the "productivity puzzle". We've had low productivity in recent years and I think this makes sense. Consumers have had stagnant or falling purchasing power and many firms can't be confident that they will be able to finance more cheaply if prices fall from current levels. (Yes I’m aware there are many possible reasons for the productivity crisis and it’s probably not any one thing. I just think this could well be one of them.)

I also think it’s important to talk about capital intensive firms. Many new industries are unusual in historical terms because they have huge initial start-up costs and then relatively low variable costs (not necessarily low but relatively low). Facebook, Google, etc. That looks less like the graph above and more like the graph below. The cost of capital and the differences between firms matter in the more traditional circumstances in the graph above but even more in the one below. Firms are not actually all identical, they’re all different and facing different circumstances. To help see that more clearly we can cheat a little and consider something more unusual, like oil.

Some oil-producing countries have access to oil deposits where the production costs per barrel of oil are low because the deposits are easy to get at, like Saudi Arabia, meanwhile some countries can only access their deposits through capital intensive and expensive processes. The US produced barely any oil until a new process (fracking) was introduced recently. That allowed it to produce oil using that process but the cost per barrel is just far higher, even now after cost savings and improvements. The price is set by the marginal barrel of oil, the last one, so if the price is high enough for US shale oil producers to make a profit then that price is also high enough to give Saudi Arabia a substantial profit margin that won’t be eroded away via competition. In real life Saudi worked with allies to create a price war to try and drive the price below what US firms could manage, that’s exactly the kind of thing you can do if you’re the player in the market that has lower costs than the marginal producers working on slim margins. That’s the idea that crosses over into everyday markets and corporations. If you have the scenario talked about above where big firms have a cost of capital advantage over small firms then they can buy them out and corner the market. This becomes even more extreme in capital intensive fields. For reasons that become obvious when you look at the graph below.

Imagine if that initial starting point on the yellow line was just a touch higher but the slope was the same. That’s the sort of situation that would face a competitor to Google that faced higher capital costs because it was riskier than the established player in the market. They’d never be able to make a profit.

The optimum profit margin on this graph, which I’ve foolishly made hard to see, is at 48 and 49 units sold, which yield identical profits. Network effects are included in consumer demand where the product isn’t very useful when not many people have it and then becomes much more useful until you get to the people who want it but are only willing to pay a small amount. (In practice obviously, Facebook and Google make their money from ads but it doesn’t affect the point I’m making here.) If the market here is 50 people in total then any competitor faces higher capital costs and the problems of network effects and a market that it's difficult to split up into smaller profitable segments because only once you have many many customers do you get into revenues high enough to cover the start-up capital costs. All of that, or any one of those factors, reduces competition and tends towards agglomeration.

Even if we’re not talking about a service with network effects but does have high capital intensity relative to the subsequent variable costs then it really does make a substantial difference what capital costs for a particular firm. Amazon, for example (which does also have network effects on top), is now the ‘Saudi Arabia’ in this example and any potential competitor is the ‘US shale industry’. That’s going to make it very difficult for markets to be truly competitive.

To wrap up, my point overall is that I’ve always found the traditional explanation of prices as firms altering variable costs to maximize profits at the price set by supply and demand to be unhelpful as a mental shortcut. Thinking about it the way I’ve laid out here is just as intuitive and, I think, just as important. Market prices won’t drop as low as they otherwise could if companies and market entrants are scared to invest. That creates rentier income for those already in the market. What this implies from a policy perspective is that inflation might well matter and it may be worth having a higher inflation target and it may also be worth having higher interest rates. A company secure in the knowledge that prices and capital costs are stable is more likely to invest. A market with nominal interest rates that aren’t too low is more competitive but low real rates maximize productive investment. If we could have a slightly high inflation target with mildly negative real rates permanently (by altering fiscal policy) I think that would be optimal in creating the stability and security necessary for firms to invest accurately and most productively.