The cost of implementing new taxes and enforcing them is usually very high. There are multiple examples of governments expending more capital trying to implementing and enforcing the tax than the amount they manage to collect. For example, from 1975 to 1986 the US net result from trying to tax capital was a loss of revenue (Gordon and Slemrod, 1988; Shoven, 1991). This creates a paradox. Theoretically, governments are supposed to tax people to make their citizens better off. However, if they use taxpayers' money to get more taxes and end up collecting less money than the amount they spent, the goal to collect such taxes is questionable for the beginning.
Capital income taxation, specifically, can be relatively easily collected by forcing national financial intermediaries (like bank and brokerages) to report their clients earning and implementing flat withholding taxation on those gains. However, governments cannot force financial intermediaries abroad to report their clients, which encourages illegal tax evasion to foreign countries. Today, the only country seriously trying to force foreign financial intermediaries to report their clients is the US through the so-called Foreign Account Tax Compliance Act (FATCA). However, by doing that the US has deprived its citizens of opening accounts in most of the best financial institutions in the world, as these institutions do not want to spend a large part of their revenue to comply with FATCA. Additionally, FATCA has spiked American renunciation rates. To the best of my knowledge, currently, no country in the world has more citizenship renunciation per year than the "land of the free." Once Americans renounce their citizenship and they no longer have any financial ties to the US, there is nothing else Uncle Sam can do to tax these individuals. Furthermore, countries abroad can hide the personal information of clients and decrease barriers for tax evaders to attract more foreign investment (Gravelle, 2009). Some scholars go as far as to say that no capital income tax will survive in an open economy (Razin and Sadka, 1991).
The fact that dividends and interest are often taxed differently and capital gains are also only taxed once gains are realized is a problem because it creates ways to legally avoid and facilitate illegal evasion. For example, investors can buy stocks in a no-income-tax jurisdiction while they are residents of a high tax state and realize those gains only after they are the fiscal residence of a no-income-tax state. In that way, they will most likely be tax-free, legally (For more check capital gains). Important studies suggest that dividends and interest should be taxed at the same rate and at accrual rather than when realized (Diamond and Mirrlees, 1971). However, financial intermediaries usually do not even have the information necessary for such a tax system to be implemented. And even if it could be possible, the taxation of non-realized investments would generate collateral damages in the whole economy (e.g. the price of all capital assets would drop and many sectors of the economy would develop slower) and raise morality questions (e.g. why should one be taxed for something they are not extracting monetary benefits?).
Several countries - such as Australia, Japan, France, Italy, and Sweden - have tried to prevent their citizens to invest abroad by implementing controls. However, capital controls are costly, hard to enforce, and discourage legit international investments (Gros, 1990; Gordon and Jun, 1993). Big economies trying to manipulate prices by controlling foreign investments also decrease the wealth of the world, on average, as it leads to a non-efficient global allocation of resources (see The "Their Tax" Fallacy). Due to the former reasons, virtually all countries have abandoned the capital controls practice. Some people also request for the so-called carryover basis of capital gains tax after death, so that the heir continues to be responsible for the appreciation of the asset, like stocks, during his ancestor life. If governments cannot even tax capital gains during one lifetime, many scholars have questioned the ability of the state to enforce a tax that can go on for theoretically infinite generations (Kopczuk, 2013).
The same major implementation and enforcement costs can be seen in many other taxes. For example, individuals and firms have avoided consumption taxes (sales tax, VAT, use tax) by buying goods online in low consumption tax jurisdictions. Countries can tax VAT at international borders but those are very costly, so the practice has been abandoned in many borders, like within the European Union (Gordon and Hines, 2002). Partially also due to low cost-benefit wealth tax and the estate tax have been eliminated by countries like Canada, Australia, and Sweden (Boadway et al, 2010; Schmalbeck, 2001). Figure 1 shows that intergenerational taxes never account for even 1 % of national GDP in the G7 economies and they are hard to enforce. Wealth tax, specifically, is very hard to be implemented. Virtually no government has decent wealth information about its citizens. Therefore, the first step to implement a wealth tax would be to collect data on the wealth of all residents, keep an annual track of that wealth, and monitor wealth transfer. Otherwise, individuals can simply move their wealth to family members (Brown, 1991). Additionally, a wealth tax would encourage wealthy individuals to reallocate their capital in items that are hard to assess value (like human capital, works of arts, rare paintings, ancient sculptures). Consequently, the wealth tax would also decrease the value of other assets in the economy that they possessed (like properties, stocks, and bonds), making most of the society worse off. Studies have questioned the capacity of wealth tax to fight income inequality (McCaffery, 1994). Not to say that it creates inefficiencies in the free-market price mechanism.