The Logic of Public Spending and Artificial Intelligence
Translated from Russian with the help of AI. Original: Russian version
The effectiveness of public spending depends on the goals set and on assessing how each decision affects the economy as a whole; AI is merely a tool.
Artificial intelligence is becoming the main driver of development today, and we, like the advanced economies, want to apply AI everywhere. But from my own experience, I can say that AI is just a tool. Everything depends on how a given process is organized and what brief the AI is given. Buying an IT solution is easy, but if we do not develop the underlying logic of business processes, we will end up throwing money to the wind.
The effectiveness of any public spending depends on the goals we set and on testing every decision against those goals through a properly organized process.
It is very important to understand the impact of any action not only on the object of improvement, but also on other objects. If a decision is meant to solve a problem for a narrow group of people but harms other participants in the economy, it is highly questionable. There are many such decisions, and I will give a few examples.
Take the decision to allow UAPF pension savings to be used for housing. We helped people buy homes who could have bought them on their own or taken out a market-rate mortgage, because they earn high salaries. But for millions of others who have no savings, buying a home has become even harder: housing prices rose sharply as a result, putting it further out of reach.
Or take tax cuts for small business. We make them so that small business will develop, GDP will grow and the share of SMEs in GDP will increase. On the one hand, the number of small enterprises has indeed grown. In practice, though, this is happening through the splitting up of medium-sized businesses in order to optimize tax payments and reduce the tax burden. On the whole, this discourages the growth of medium-sized businesses and makes them opaque, which affects their creditworthiness, since opaque companies are harder to assess. As a result, we are not getting new large companies in the market, yet the foundation of any economy is, after all, stable medium-sized and large enterprises. A group of small enterprises is poorly managed and hard to grow. This has also led to a decline in tax revenue (look at it as a percentage of GDP), and thereby to a budget deficit and a heavier tax burden on large enterprises.
The same applies to interest rate subsidies, various tax breaks, the construction of buildings that are not really needed, and the development of some sectors by imposing demand on others, to the detriment of their development, and so on.
Every investment should be evaluated solely against its end goals, through the lens of the "five whys". For that, artificial intelligence is not always needed; ordinary human intelligence is enough.
It seems to me that we need to focus on decision-making processes first, and only then introduce automation tools. Every government body needs to re-engineer its decision-making. Setting the right goal, assessing risks and monitoring implementation should be at the core of changing these processes.