The rise of GPUs and AI promises significant reductions in risk and interest rates across the economy. However, the true benefits hinge on aligning fiscal, monetary, and regulatory policies with this new reality. Without proper alignment, we risk a permanent transfer of wealth within the financial sector, where lending could become expensive and disproportionately benefit large, low-risk tech companies, further concentrating profits.
Recurring returns are the lifeblood of the financial industry. They allow companies to be valued based on the net present value of their future earnings, which becomes their equity.
Consider the traditional credit system taught in universities. You need a new roof, so you approach a bank. But the bank, being a prudent lender, has already allocated its available funds. It turns to the central bank, borrowing at the overnight rate, using your roof loan as an asset to cover the central bank's rate.
In reality, banks often trade these loans, packaged as commercial paper, in large denominations (typically $1 million or more) to facilitate statistical analysis. Diversification is key: spreading risk across thousands of loans allows for accurate measurement of default rates. Institutions like Fannie Mae and Freddie Mac historically secured these rates by purchasing mortgage-backed securities from banks.
Banks thrive on diversification. The smaller the loans and the more customers they have, the better they do. A diverse customer base allows them to offset individual failures with successes. Systemic risks are their biggest threat. Overnight borrowing is a last resort, viewed by many conservative banks as a potential source of losses and inflation. The trust of depositors is superficial.
Ideally, the overnight federal rate should cover the central bank's risks, with the bank's margin covering the specific risks of its loan portfolio. Citizens could borrow directly from the central bank, eliminating the intermediary. However, centralized lending can be bogged down by bureaucracy and risk. Interest rates, in effect, act as a safeguard, incentivizing property owners to collaborate and reduce their collective risk. AI has the potential to further mitigate these risks through enhanced communication, identification and elimination.
Countries actually choose different measures to drive their currency. Researchers can replicate this by issuing a cryptocurrency for games and setting their policies. The German tradition followed by the EU used to be protecting the buying power of the money by manipulating exchange rates. Keeping unemployment at bay and the economy working is the usual American approach. The Japanese system embraces accounting and that every part of the system gets money to run with. Accumulated savings is the side effect of such policies. Japan has both high savings, but high government debt as well. Saudi Arabia empowered its citizens with cash backed by oil reserves. Researchers can easily experiment nowadays with cryptocurrencies that track time spent on work or volunteering. Similarly, decentralized cryptocurrencies that allow raising a capped amount every month by every citizen have an interesting future.
Ultimately, risks and rates should converge. Divergence leads to imbalances, effectively taxing consumption to benefit either asset holders or borrowers, hindering growth and potentially triggering recession. Divergence depletes the assets on one side, eliminating their market. Look at the diagram as an example.
Imagine large sovereign wealth funds, like those in Saudi Arabia. With access to vast data, they can fine-tune their fees based on loan performance. Being too generous risks bankruptcy, while being too greedy can drive clients out of business. A sustainable system requires rates and risks to align, favoring either large financial institutions or a multitude of small investors equipped with tools like Robinhood.
The lowest risks and rates are found in economies built on annual recurring revenues. This predictability provides statistical evidence of a business's future viability, simplifying decision-making. This creates a closed-loop system with two key characteristics:
The sum of inflows and outflows at each point equals zero. The demand and supply of labor for these inflows and outflows also equal zero.
Think of carrying a dollar from a valley to a summit and back down via a different route. The effort invested uphill generates goods downhill. Spending may slow if you save and choose a less steep route. The cumulative elevation gain on each circular route is zero.
These principles mirror Kirchhoff's circuit laws in electrical engineering. You can change paths at intersections, altering your savings or earnings strategies. These laws also account for savings (like batteries) and debt (like capacitors).
There's a catch: people can get "stuck in the parking lot," akin to a battery effect. Saving is beneficial if central banks effectively manage inflows and outflows.
Savings drive retirement consumption and risk reduction for the median citizen. The buying power is secured if it is invested in the production that will eventually use it. Recessions oftentimes trigger fear to change it into something with a fixed amount like gold or crypto. This shows at least an illusion of power on new graduates to pay doctors or caregivers. Oftentimes just a portfolio balanced enough to the new economy reduces risks enough.
Fiscal politics is an additional invention to keep the economy running by adding money at certain points. This flow can be beneficial, but the indicator called multiplier effect measures its efficiency.
Not every dollar has the same impact. Multiplier effects are biggest when the initial government grants are spent immediately and circulate in the economy as recurring revenue. If it ends up as savings to boost the bank's deposits to lend more, it empowers the already rich, or creates factories already in supply, which can mitigate the effect. Fiscal politics also changes when governing parties or leaders change. This reduces the recurring effect.
The multiplier effects, analogous to reduced loss in machine learning, or additional voltage in Kirchhoff's laws, are sometimes a better but harder-to-calculate measure than interest rates, inflation, or exchange rates. Still, they are closely related, since all of these are related to price-to-earnings ratios and time to pay off investments.