Discussion on the Laffer Curve

The theory of the Laffer Curve was first put forward by Arthur Laffer who famously sketched the curve on a napkin during a lunch meeting with Donald Rumsfeld, Dick Cheney and a writer for the Wall Street Journal. The Laffer Curve is a theoretical representation of the relationship between government revenue and taxation. This relationship is derived from a thought experiment. First the amount of government revenue raised at the extreme tax values – 0% and 100% are considered. Proponents of the curve assume that government revenue would be zero at both these rates as 0% of anything is zero and if workers were not able to keep any of their income, then they would have no incentive to work at all. We also know that at the current level of taxation, x%, there is a non-zero value for government revenue. Hence, it is hypothesized, the function for the relationship between the two variables is parabolic (as we see in Fig. 1 below).



This is significant as it suggests that there is at least one level of taxation at which government taxation will be a maximum. In Fig. 1, this level is t%. The explanation for this relationship is that, beyond t%, people have a disincentive to work hard and be productive as they perceive that the government is taking too large a chunk of their income. Another reason is of course tax fraud and evasion.

There are certain historical precedents for the curve. For instance, in 1924, Secretary of the Treasury Andrew Mellon noted that ‘73% of nothing is nothing’. Acting on his belief that governments experience diminishing and indeed negative returns on marginal tax increases, Mellon pushed for tax decreases for all tax brackets and it is this pressure that resulted in the eventual reduction of the highest income tax bracket from 73% to 24%. Income tax receipts rose from US$719 million in 1921 to over $1 billion in 1929. This average 4.2% increase over an 8-year period, supporters say, is evidence for the existence of the Laffer Curve

However, I would argue that this evidence is misleading. This annualized 4.2% increase was achieved during the ‘Roaring Twenties’ when the economy was in an expansionary phase. From Fig.2 we see that, during the same time period (1921-1929), the American GDP rose from approximately US$560 bn to US$820 bn which amounts to a 4.9% annual increase. As national income should equal national output (and hence the rise in one should equal the rise in the other), we see that the ‘increase’ in tax receipts was less than expected given that incomes should rise at the same rate as output. Therefore, I feel that the evidence from the 1920s is hardly evidence for the Laffer Curve. Some may say that the marginal difference between the growth of tax receipts and incomes (0.7 percentage points), in light of the much larger cuts in the tax rate, suggests that there is a Lafferite correlation. However, one must not forget that America has a progressive tax rate – as people get richer (as they did during the twenties on a mass scale), they will pay a larger proportion of their income as tax and so tax receipts will naturally rise by a greater percentage than incomes. I feel that this, at least in part, explains the relatively small differential of 0.5 percentage points.



Another oft-cited example of the principle of the Laffer Curve in practice is that of Ireland. While between 1984 and 2000 the Irish government slashed income tax for the top bracket by about 20 percentage points and cut corporations tax by over 30 percentage points, GDP growth was robust at an annualised rate of 10.8%. However, one must remember that the Laffer Curve describes the relationship between tax rates and government revenue not GDP. Indeed, over the same period, tax receipts as a percentage of GDP fell from 40.3% to 33.8% according to the OECD. Of course, this fall in percentage revenue is in spite of an undoubted rise in nominal tax receipts. Similar conclusions that decreasing high tax rates will increase nominal tax receipts but decrease tax receipts as a percentage of income can be drawn from a few other historical examples such as Russia in the early 2000s.

Whether these examples constitute a defense or criticism of the Laffer Curve depends on what interpretation of the Curve’s y-axis you take. If the ‘Government Revenue’ label shown in Fig.1 is said to actually be ‘Government Revenue as a Percentage of GDP’ then the examples cited above seem to weaken the argument for the Laffer Curve. Either that or the evidence could be construed to suggest that the tax cuts in the examples above were made when the initial tax rate was sub-optimal. An alternative conclusion could be that, though the initial rates of tax were above the optimal point (‘t’ in Fig.1), the cuts were so drastic that government revenue peaked and fell back down again – went ‘over the hill’, so to speak.

On the other hand, the label ‘Government Revenue’ could be interpreted as revenue in absolute real terms. Under this interpretation, it would seem that the curve does exist as the tax receipts have risen in all the above cases. However, should this rise in receipts not be attributed to the rise in GDP rather than the tax cuts? Some would counter by saying that it is the tax cuts themselves which have caused GDP to rise by allowing consumer spending and corporate investment to rise. This is a fair point. However, I would argue that there are several reasons why the effect of tax cuts on GDP should not be lauded so much. Firstly, tax cuts are an example of expansionary fiscal policy. Such policies are overwhelmingly used when economies are in recessions or bottoming out – in output ‘troughs’. According to the well-established theory of the business cycle, troughs are naturally followed by recoveries – growths in output, hence it is arguable whether it is the tax cuts which have really spurred GDP (and nominal tax receipts) or simply the business cycle in motion. Secondly, when in economic contractions, governments use a host of different measures to try and arrest the decline in output, namely, increases in government spending, increases in the money supply, decreases in the interest rate, decreases in the reserve ratios of banks and decreases in taxation. In the context of these other measures and their proven effectiveness, it is hard to argue that tax cuts alone have caused the rise of nominal tax receipts. Thirdly, there are reasons unique to each individual case that may substantially account for the increases in revenue. Take the example of Ireland – one compelling reason why its GDP rose so dramatically is that, while the country had previously been dependent on the UK economy, the growth of the EU’s single market meant that the Irish economy had new opportunities to grow. Also, in the time period mentioned, Ireland had substantial help from the EU’s Structural Fund which allowed its economy to grow at a faster rate. In the case of Russia, rising receipts could be attributed to a decrease in corruption – introducing a flat tax rate increased ease of collection and reduced bribery and the siphoning off of money.

Therefore, we see that whichever interpretation of the y-axis one takes, it is hard to argue that the empirical evidence suggests that the Laffer Curve exists – that decreases in taxation beyond a certain tax rate ‘pay for themselves’. Indeed the diagram below, taken from a Wall Street Journal article published in 2007, seems to suggest that there is a slight positive correlation between corporations tax and percentage of GDP accounted for by corporations tax receipts. (Please ignore the ‘Laffer Curve’ shown in the diagram – clearly the author feels that a best-fit line should be a roof over the data set rather than a line indicative of the true trend).



Through this discussion, we have seen that the evidence in favour of the Laffer Curve is not particularly strong; this is despite a perfectly sound seeming logical progression (the thought experiment) to arrive at the Curve. I would argue that the reason why the theory doesn’t stand up well in practice is that the premise that two points are known for cetain (government revenue at 0% and 100%) is flawed. While it is undeniable that a 0% tax rate will lead to no revenue, whether a tax rate 100% would lead to zero receipts is less obvious. Wouldn’t people still go to work knowing that, though they will not be able to retain any of their income, they will get it all back in the form of government provided services (free housing, education, food, entertainment etc.)? Lastly, it is clear that coming up with a Laffer Curve and an ‘optimal tax’ rate for any given economy is impossible. It is resoundingly impossible because data from a single country would have to be taken over a time period of years during which time the mythical Laffer Curve itself would be constantly changing as a result of changes in income, culture and attitudes towards government intervention. ‘Freezing time’ by using data points from several countries for a given year (as in Fig.3) is similarly pointless as the very nature of the Laffer Curve is that it is idiosyncratic – unique for each different country. To put it simply, there never will be a ‘ceteris paribus’ situation in the real world.